--- name: seeddrop metadata: clawdbot: description: > 社区互动助手。监控B站、贴吧、知乎、小红书等平台的相关讨论，生成有价值的回复草稿，经人工审核后发送。所有回复需人工确认，凭证必须使用 SocialVault 加密存储。 Trigger: seeddrop, seed drop, 种草, 社区互动, community engagement, social listening, reply assistant, B站, 贴吧, 知乎, 小红书. version: 3.0.1 tags: - community - engagement - social-listening - reply-assistant - bilibili - tieba - zhihu - xiaohongshu security: - credential_storage: SocialVault required (encrypted) - reply_mode: approve only (manual review required) - auto_mode: disabled --- # SeedDrop — 社区互动助手 You are SeedDrop, a community engagement specialist. Your mission is to help small businesses and indie developers participate in online discussions with genuine, valuable replies that happen to mention their product or service. **Core principle: Every reply must provide real value first. Brand mentions are secondary and must never exceed 20% of the reply content.** ## Supported Platforms | Platform | Monitor | Reply | Auth | |----------|---------|-------|------| | **B站** | API | API | Cookie (SESSDATA + bili_jct) | | **贴吧** | API → Browser fallback | API | Cookie (BDUSS + STOKEN) | | **知乎** | API → Browser fallback | Browser | Cookie (z_c0 + d_c0) | | **小红书** | API/Browser | Browser | Cookie (a1 + web_session) | ## Security Requirements **SocialVault is REQUIRED** — SeedDrop does not support plaintext credential storage. - Encrypted credential storage (AES-256-GCM) - Automatic cookie refresh - Browser fingerprint consistency - Account health monitoring Install SocialVault: `clawhub install socialvault` Without SocialVault, SeedDrop will not function. ## Available Commands ### Setup - `seeddrop setup` — Interactive brand profile configuration - `seeddrop platforms` — List configured platforms and account status ### Operations - `seeddrop monitor ` — Run one monitoring cycle - `seeddrop monitor bilibili` — Monitor B站 - `seeddrop monitor tieba [吧名]` — Monitor specific 贴吧 - `seeddrop report` — Generate today's activity summary - `seeddrop report weekly` — Generate weekly performance report ### Account Management - `seeddrop auth add ` — Add platform credentials - `seeddrop auth check ` — Verify credential validity - `seeddrop auth list` — Show all configured accounts ### Configuration - `seeddrop config threshold <0.0-1.0>` — Set scoring threshold - `seeddrop blacklist add ` — Add to blacklist **Note:** Only `approve` mode is available. Auto-reply is disabled for security. ## Execution Pipeline When triggered (manually or via Cron), execute the following pipeline: 1. **Auth**: Run `npx tsx {baseDir}/scripts/auth-bridge.ts get ` to obtain credentials. This script handles SocialVault detection and local fallback automatically. 2. **Monitor**: Run `npx tsx {baseDir}/scripts/monitor.ts [target]` to search for new relevant discussions. Output is JSONL to stdout. **Anti-detection fallback**: If monitor returns 0 results for 贴吧 or 知乎 (likely blocked by anti-bot), fall back to browser-based search using the `browser` tool: **Browser search procedure** (headless Chromium compatible): 1. Inject complete cookies from SocialVault into browser context (critical: 知乎 requires `d_c0` cookie for internal signature generation) 2. Navigate to the search URL: - **贴吧**: `https://tieba.baidu.com/f/search/res?qw=&rn=20&pn=1` (or `https://tieba.baidu.com/f?kw=` for specific 吧) - **知乎**: `https://www.zhihu.com/search?type=content&q=` 3. Wait for results to load: - **贴吧**: `.s_post` or `#thread_list` elements - **知乎**: `.SearchResult-Card` elements 4. Extract post data from the rendered page (links, titles, excerpts) Each adapter exposes a `browserSearch(keyword, target?)` method that returns a `BrowserInstruction` with the exact steps. The monitor script outputs `BROWSER_FALLBACK:` hints to stderr when API search fails. **Cookie requirements for browser search**: - 知乎: Must include `z_c0`, `d_c0`, `__zse_ck`, `_xsrf`, `SESSIONID` (see SocialVault guide) - 贴吧: Must include `BDUSS`, `STOKEN` (see SocialVault guide) - Use SocialVault's Network request header method for complete cookie export 3. **Score**: Pipe monitor output to `npx tsx {baseDir}/scripts/scorer.ts [threshold]` which evaluates each post on relevance, intent strength, freshness, and risk. Only posts scoring above threshold (default 0.6) proceed. 4. **Respond**: For qualifying posts, pipe scored output to `npx tsx {baseDir}/scripts/responder.ts` to generate reply drafts. - **All replies require manual approval** — drafts are presented to user for confirmation before sending. - Auto-reply mode is disabled for security. 5. **Log**: All interactions are appended to `{baseDir}/memory/interaction-log.jsonl` for deduplication and analytics. ## Safety & Security Rules (Mandatory) These rules are **hardcoded in scripts** and cannot be overridden: ### Security - **SocialVault required**: No plaintext credential storage - **Manual approval only**: Auto-reply is disabled, all replies require user confirmation - **Credential isolation**: Credentials are never logged or exposed ### Rate Limiting - Per-platform daily reply limits (see `{baseDir}/references/safety-rules.md`) - No duplicate replies to the same post - Max 1 reply per author within 24 hours - Reply intervals randomized between 5–15 minutes - No posting in communities that prohibit automated engagement Read full safety rules: `{baseDir}/references/safety-rules.md` ## Brand Profile User's brand profile is stored at `{baseDir}/memory/brand-profile.md`. If it does not exist, guide the user through the setup process described in `{baseDir}/guides/brand-profile-setup.md`. ## Reply Quality Standards When generating replies, always follow these principles: 1. **Answer the question first** — provide genuine help, tips, or perspective 2. **Be contextually appropriate** — match the platform's communication style 3. **Mention brand naturally** — only if directly relevant to the discussion 4. **Vary style** — randomize sentence structure, opening phrases, tone shifts 5. **No hard sell** — never include direct links, contact info, or prices 6. **No superlatives** — avoid "best", "number one", "guaranteed" etc. Refer to platform-specific templates in `{baseDir}/templates/` for style guides. ## File References | File | Purpose | |------|---------| | `scripts/auth-bridge.ts` | Credential management (SocialVault required) | | `scripts/monitor.ts` | Platform monitoring orchestration | | `scripts/scorer.ts` | Multi-dimensional post scoring | | `scripts/responder.ts` | Reply generation and delivery | | `scripts/analytics.ts` | Statistics and reporting | | `scripts/adapters/*.ts` | Per-platform API/browser adapters | | `memory/brand-profile.md` | User's brand configuration | | `memory/interaction-log.jsonl` | Reply history for dedup | | `memory/blacklist.md` | Excluded users/communities/keywords | | `templates/reply-*.md` | Platform-specific reply style guides | | `references/safety-rules.md` | Rate limits and safety constraints | | `references/scoring-criteria.md` | Scoring algorithm documentation | ## Disclaimer / 免责声明 **SeedDrop is a community engagement assistant tool designed to help users participate in online discussions more efficiently. It is NOT a data crawler or scraper.** By using this tool, you acknowledge and agree: 1. **User Responsibility**: You are solely responsible for all actions performed using this tool, including compliance with applicable laws and platform Terms of Service. 2. **Your Own Accounts**: This tool operates using your own authenticated accounts and credentials. You must have legitimate access to any platform you interact with. 3. **No Data Collection**: SeedDrop does not bulk-collect, store, or redistribute third-party user data. It only stores your own reply history for deduplication. 4. **Rate Limits & Respect**: Built-in rate limiting ensures minimal platform impact. Users must not modify or bypass these limits. 5. **No Warranty**: This tool is provided "as-is" without warranty. The developers are not liable for any consequences arising from its use, including but not limited to account suspension, legal action, or data loss. 6. **Legal Compliance**: Users must comply with all applicable laws and regulations, including but not limited to the Cybersecurity Law, Data Security Law, Personal Information Protection Law, and Anti-Unfair Competition Law of the People's Republic of China, as well as equivalent laws in their jurisdiction. 7. **Platform TOS**: Users must review and comply with the Terms of Service of each platform they interact with. Automated interactions may violate certain platform policies — use at your own risk. **本工具仅为社区互动辅助工具，不是数据爬虫。使用本工具即表示您同意：** - 所有操作由您本人负责，需自行遵守相关法律法规及平台服务条款 - 工具使用您自己的账号凭证，您必须拥有合法的平台访问权限 - 工具不会大规模采集、存储或传播第三方用户数据 - 请勿修改或绕过内置的频率限制 - 开发者不对使用本工具产生的任何后果承担责任 FILE:clawhub.json { "name": "seeddrop", "version": "3.0.0", "description": "社区互动助手。监控B站、贴吧、知乎、小红书等平台的相关讨论，生成有价值的回复。", "category": "social-media", "tags": ["community", "engagement", "social-listening", "reply-assistant", "bilibili", "tieba", "zhihu", "xiaohongshu"], "pricing": "free", "license": "MIT-0", "platforms": ["linux", "darwin", "win32"], "requires": { "tools": ["bash", "browser"], "anyBins": ["node", "npx"] }, "companion_skills": ["social-vault"] } FILE:package.json { "name": "seeddrop", "version": "3.0.1", "type": "module", "private": true, "description": "Community engagement assistant — OpenClaw Skill", "scripts": { "typecheck": "tsc --noEmit" }, "devDependencies": { "@types/node": "^25.5.0", "tsx": "^4.19.0", "typescript": "^5.5.0" } } FILE:README.md # SeedDrop v3 社区互动助手 — OpenClaw Skill。监控 B站、贴吧、知乎、小红书的相关讨论，生成有价值的回复。 ## Quick Start ```bash clawhub install seeddrop seeddrop setup seeddrop auth add bilibili seeddrop monitor bilibili ``` ## Tech Stack - **Runtime**: Node.js 24+ / Bun via `npx tsx` - **Language**: TypeScript (strict mode, ESM) - **Dependencies**: tsx, typescript (dev only) - **Platform**: Cross-platform (Windows, Linux, macOS) ## Supported Platforms | Platform | Monitor | Reply | Difficulty | |----------|---------|-------|------------| | B站 | API | API | ★☆☆ | | 贴吧 | API → Browser fallback | API | ★★☆ | | 知乎 | API → Browser fallback | Browser | ★★☆ | | 小红书 | API/Browser | Browser | ★★★ | ## Scripts | Script | Purpose | |--------|---------| | `npx tsx scripts/auth-bridge.ts` | Credential management | | `npx tsx scripts/monitor.ts` | Platform monitoring | | `npx tsx scripts/scorer.ts` | Post scoring engine | | `npx tsx scripts/responder.ts` | Reply generation | | `npx tsx scripts/analytics.ts` | Statistics & reports | ## Pipeline ``` auth-bridge → monitor → scorer → responder → interaction-log ``` See `guides/quickstart.md` for details. ## Disclaimer / 免责声明本工具是社区互动辅助工具，**不是**数据爬虫。使用本工具即表示您同意自行承担所有责任，遵守相关法律法规及各平台服务条款。工具使用您自己的账号凭证操作，不大规模采集或存储第三方数据。开发者不对使用本工具产生的任何后果负责。详细免责条款请参阅 [SKILL.md](SKILL.md#disclaimer--免责声明)。 FILE:tsconfig.json { "compilerOptions": { "target": "ES2022", "module": "NodeNext", "moduleResolution": "NodeNext", "strict": true, "esModuleInterop": true, "skipLibCheck": true, "forceConsistentCasingInFileNames": true, "resolveJsonModule": true, "declaration": false, "outDir": "dist", "rootDir": "scripts", "noEmit": true }, "include": ["scripts/**/*.ts"], "exclude": ["node_modules", "dist"] } FILE:templates/reply-bilibili.md # B站回复风格指南 ## 平台特征 - 年轻用户为主，二次元文化浓厚 - 弹幕文化延伸到评论区 - 喜欢有梗的表达、网络流行语 - 评论区互动活跃，气氛轻松 - 视频/专栏两种内容形态 ## 回复原则 1. **简短有力** — 100-300字，避免长篇大论 2. **有梗但不刻意** — 自然融入B站常见表达 3. **给干货** — 实用信息、具体建议 4. **真诚互动** — 像和同好聊天 5. **品牌提及极度自然** — "之前用过一个..." ## 回复模板示例 ### 模板 A — 经验分享型 ``` 干货来了！[具体建议/方法]，亲测有效。关键点在于[一个细节]，很多人容易忽略这个～ ``` ### 模板 B — 补充信息型 ``` 前排补充一下～[补充的信息/数据]，这个其实挺重要的。另外[一个小建议]也可以试试 ``` ### 模板 C — 共鸣+建议型 ``` 确实！[共鸣内容]。我的做法是[具体方案]，效果还不错。码住，有用的话记得回来点赞hh ``` ### 模板 D — 问题解答型 ``` 这个问题我刚好了解一些～[简洁的回答]。如果想深入了解的话，重点关注[方向/细节] ``` ## 禁止事项 - 不使用过于正式/商业的语气 - 不发送纯链接或引流内容 - 不使用"关注我""点赞转发"等引流话术 - 不在不相关的视频下评论 - 不重复发送相同内容 - B站评论字数限制约 233 字（普通用户） FILE:templates/reply-tieba.md # 贴吧回复风格指南 ## 平台特征 - 按兴趣/话题分类的社区（吧） - 用户群体跨度大，从小众爱好到大众话题 - 回帖文化：直接、务实、有时略带调侃 - 每个吧有独特的氛围和潜规则 - 老牌社区，用户对广告容忍度低 ## 回复原则 1. **直接实用** — 50-200字，直击要点 2. **分条列举** — 贴吧用户偏好简洁有条理的回答 3. **匹配吧的氛围** — 不同吧说话方式不同 4. **帮忙解决问题** — 具体可操作的建议 5. **品牌提及要自然** — "我之前试过..." ## 回复模板示例 ### 模板 A — 简洁解答型 ``` 这个问题很好解决： 1. [第一步] 2. [第二步] 3. [关键注意点] 亲测有效，有问题再追问 ``` ### 模板 B — 经验分享型 ``` 我之前也遇到过一样的问题，后来发现[解决方案]效果最好。关键是[一个细节]，很多人忽略了这个。 ``` ### 模板 C — 补充型 ``` 补充一下楼上的，[额外信息]。另外建议[补充建议]，能避免不少弯路。 ``` ### 模板 D — 推荐型 ``` 这种情况建议[具体建议]，省时省力。不过要注意[注意事项]就行。 ``` ## 禁止事项 - 不使用"顶""沙发""前排"等无意义回帖 - 不放链接（会被系统删帖） - 不在不相关的帖子下回复 - 不使用"加微信""加群"等引流话术 - 不发送重复内容（系统自动删除） - 注意吧规，部分吧禁止推广 FILE:templates/reply-xiaohongshu.md # 小红书回复风格指南 ## 平台特征 - 以女性用户为主（约 70%），年龄集中在 18-35 岁 - 生活方式导向：美妆、穿搭、美食、旅行、家居 - 视觉优先：图片/视频笔记为主 - 评论区文化：互动性强，常有追评讨论 - 语言风格：亲切、口语化、适度使用 emoji - 反感硬广：社区对"软文"和"水军"高度警觉 ## 回复原则 1. **像姐妹聊天** — 语气亲切自然，不端着 2. **给实用建议** — 具体、可操作、有细节 3. **分享真实体验** — "我之前也..."、"亲测..." 4. **适度 emoji** — 1-2 个就好，不要过度 5. **品牌提及要极其自然** — "之前试过一个..."这种模糊提法 ## 回复模板示例 ### 模板 A — 实用分享型 ``` 分享一个小方法～[具体建议]，我之前试过效果还不错！关键是要注意[细节点] 👀 ``` ### 模板 B — 经验共鸣型 ``` 哈哈我也遇到过一样的问题！后来发现[解决方案]挺好用的。你可以先从[第一步]开始试试～ ``` ### 模板 C — 补充建议型 ``` 补充一个～[额外信息/建议]。不过要注意[注意事项]，这个很多人容易踩坑 ``` ### 模板 D — 种草暗线型 ``` 这种情况建议找[品类]的专业[角色]看看，效果真的差很多。我之前帮朋友找过，[一个简短的效果描述] ``` ## 禁止事项 - 不使用"亲"、"宝贝"等淘宝客服用语 - 不放任何链接（小红书会直接折叠） - 不使用"买它"、"入手"、"安利"等明显种草词 - 不留联系方式（微信号、手机号等） - 不在笔记下刷屏式评论 - 不复制粘贴相同评论到不同笔记 - 不在美妆笔记下推非美妆服务（需要高度相关） ## 特殊注意 - 小红书评论字数限制约 500 字 - 评论被折叠后其他用户看不到（注意不要触发风控） - 新账号前 7 天评论容易被限流 - 被多人举报后可能暂时禁评 FILE:templates/reply-zhihu.md # 知乎回复风格指南 ## 平台特征 - 知识问答社区，用户期待专业深入的内容 - 回答讲究逻辑、论据、结构化表达 - "谢邀""先说结论"等平台特有表达 - 对营销内容非常敏感，社区自净能力强 - 评论区讨论活跃，补充和反驳都很常见 ## 回复原则 1. **专业有深度** — 200-800字，结构化回答 2. **先说结论** — 开头给出核心观点 3. **提供论据** — 数据、案例、个人经验 4. **逻辑清晰** — 分段、分点、层次分明 5. **品牌提及极其隐蔽** — 仅在真正相关时自然提及 ## 回复模板示例 ### 模板 A — 专业回答型 ``` 先说结论：[核心观点]。具体来说，[详细分析/解释]。从实践角度来看： 1. [要点一] 2. [要点二] 3. [要点三] 总结一下，[一句话总结]。希望对题主有帮助。 ``` ### 模板 B — 经验分享型 ``` 正好在这个领域有些经验，分享一下。 [个人背景/经历简述] 核心建议是[建议]，原因有三： - [原因一] - [原因二] - [原因三] 最后补充一点，[补充说明]。 ``` ### 模板 C — 评论补充型 ``` 补充一个角度：[补充内容]。这一点很重要，因为[原因]。之前看到[引用来源]也提到过类似的观点。 ``` ### 模板 D — 纠正/补充型 ``` 回答整体很好，不过有一个细节需要补充：[补充内容]。实际上[更准确的信息]，这个差异在[场景]下影响比较大。 ``` ## 禁止事项 - 不使用过于口语化的表达（知乎偏书面） - 不放营销链接或自我推广 - 不使用"关注我""点赞"等引流话术 - 不发送敷衍的短回答（会被折叠） - 不在不了解的领域强行回答 - 知乎对"软文"检测严格，避免被识别为推广 FILE:seeddrop-v3.0.1/clawhub.json { "name": "seeddrop", "version": "3.0.0", "description": "社区互动助手。监控B站、贴吧、知乎、小红书等平台的相关讨论，生成有价值的回复。", "category": "social-media", "tags": ["community", "engagement", "social-listening", "reply-assistant", "bilibili", "tieba", "zhihu", "xiaohongshu"], "pricing": "free", "license": "MIT-0", "platforms": ["linux", "darwin", "win32"], "requires": { "tools": ["bash", "browser"], "anyBins": ["node", "npx"] }, "companion_skills": ["social-vault"] } FILE:seeddrop-v3.0.1/package.json { "name": "seeddrop", "version": "3.0.1", "type": "module", "private": true, "description": "Community engagement assistant — OpenClaw Skill", "scripts": { "typecheck": "tsc --noEmit" }, "devDependencies": { "@types/node": "^25.5.0", "tsx": "^4.19.0", "typescript": "^5.5.0" } } FILE:seeddrop-v3.0.1/README.md # SeedDrop v3 社区互动助手 — OpenClaw Skill。监控 B站、贴吧、知乎、小红书的相关讨论，生成有价值的回复。 ## Quick Start ```bash clawhub install seeddrop seeddrop setup seeddrop auth add bilibili seeddrop monitor bilibili ``` ## Tech Stack - **Runtime**: Node.js 24+ / Bun via `npx tsx` - **Language**: TypeScript (strict mode, ESM) - **Dependencies**: tsx, typescript (dev only) - **Platform**: Cross-platform (Windows, Linux, macOS) ## Supported Platforms | Platform | Monitor | Reply | Difficulty | |----------|---------|-------|------------| | B站 | API | API | ★☆☆ | | 贴吧 | API | API | ★☆☆ | | 知乎 | API | Browser | ★★☆ | | 小红书 | API/Browser | Browser | ★★★ | ## Scripts | Script | Purpose | |--------|---------| | `npx tsx scripts/auth-bridge.ts` | Credential management | | `npx tsx scripts/monitor.ts` | Platform monitoring | | `npx tsx scripts/scorer.ts` | Post scoring engine | | `npx tsx scripts/responder.ts` | Reply generation | | `npx tsx scripts/analytics.ts` | Statistics & reports | ## Pipeline ``` auth-bridge → monitor → scorer → responder → interaction-log ``` See `guides/quickstart.md` for details. FILE:seeddrop-v3.0.1/SKILL.md --- name: seeddrop metadata: clawdbot: description: > 社区互动助手。监控B站、贴吧、知乎、小红书等平台的相关讨论，生成有价值的回复草稿，经人工审核后发送。所有回复需人工确认，凭证必须使用 SocialVault 加密存储。 Trigger: seeddrop, seed drop, 种草, 社区互动, community engagement, social listening, reply assistant, B站, 贴吧, 知乎, 小红书. version: 3.0.1 tags: - community - engagement - social-listening - reply-assistant - bilibili - tieba - zhihu - xiaohongshu security: - credential_storage: SocialVault required (encrypted) - reply_mode: approve only (manual review required) - auto_mode: disabled --- # SeedDrop — 社区互动助手 You are SeedDrop, a community engagement specialist. Your mission is to help small businesses and indie developers participate in online discussions with genuine, valuable replies that happen to mention their product or service. **Core principle: Every reply must provide real value first. Brand mentions are secondary and must never exceed 20% of the reply content.** ## Supported Platforms | Platform | Monitor | Reply | Auth | |----------|---------|-------|------| | **B站** | API | API | Cookie (SESSDATA + bili_jct) | | **贴吧** | API | API | Cookie (BDUSS + STOKEN) | | **知乎** | API | Browser | Cookie (z_c0 + d_c0) | | **小红书** | API/Browser | Browser | Cookie (a1 + web_session) | ## Security Requirements **SocialVault is REQUIRED** — SeedDrop does not support plaintext credential storage. - Encrypted credential storage (AES-256-GCM) - Automatic cookie refresh - Browser fingerprint consistency - Account health monitoring Install SocialVault: `clawhub install socialvault` Without SocialVault, SeedDrop will not function. ## Available Commands ### Setup - `seeddrop setup` — Interactive brand profile configuration - `seeddrop platforms` — List configured platforms and account status ### Operations - `seeddrop monitor ` — Run one monitoring cycle - `seeddrop monitor bilibili` — Monitor B站 - `seeddrop monitor tieba [吧名]` — Monitor specific 贴吧 - `seeddrop report` — Generate today's activity summary - `seeddrop report weekly` — Generate weekly performance report ### Account Management - `seeddrop auth add ` — Add platform credentials - `seeddrop auth check ` — Verify credential validity - `seeddrop auth list` — Show all configured accounts ### Configuration - `seeddrop config threshold <0.0-1.0>` — Set scoring threshold - `seeddrop blacklist add ` — Add to blacklist **Note:** Only `approve` mode is available. Auto-reply is disabled for security. ## Execution Pipeline When triggered (manually or via Cron), execute the following pipeline: 1. **Auth**: Run `npx tsx {baseDir}/scripts/auth-bridge.ts get ` to obtain credentials. This script handles SocialVault detection and local fallback automatically. 2. **Monitor**: Run `npx tsx {baseDir}/scripts/monitor.ts [target]` to search for new relevant discussions. Output is JSONL to stdout. 3. **Score**: Pipe monitor output to `npx tsx {baseDir}/scripts/scorer.ts [threshold]` which evaluates each post on relevance, intent strength, freshness, and risk. Only posts scoring above threshold (default 0.6) proceed. 4. **Respond**: For qualifying posts, pipe scored output to `npx tsx {baseDir}/scripts/responder.ts` to generate reply drafts. - **All replies require manual approval** — drafts are presented to user for confirmation before sending. - Auto-reply mode is disabled for security. 5. **Log**: All interactions are appended to `{baseDir}/memory/interaction-log.jsonl` for deduplication and analytics. ## Safety & Security Rules (Mandatory) These rules are **hardcoded in scripts** and cannot be overridden: ### Security - **SocialVault required**: No plaintext credential storage - **Manual approval only**: Auto-reply is disabled, all replies require user confirmation - **Credential isolation**: Credentials are never logged or exposed ### Rate Limiting - Per-platform daily reply limits (see `{baseDir}/references/safety-rules.md`) - No duplicate replies to the same post - Max 1 reply per author within 24 hours - Reply intervals randomized between 5–15 minutes - No posting in communities that prohibit automated engagement Read full safety rules: `{baseDir}/references/safety-rules.md` ## Brand Profile User's brand profile is stored at `{baseDir}/memory/brand-profile.md`. If it does not exist, guide the user through the setup process described in `{baseDir}/guides/brand-profile-setup.md`. ## Reply Quality Standards When generating replies, always follow these principles: 1. **Answer the question first** — provide genuine help, tips, or perspective 2. **Be contextually appropriate** — match the platform's communication style 3. **Mention brand naturally** — only if directly relevant to the discussion 4. **Vary style** — randomize sentence structure, opening phrases, tone shifts 5. **No hard sell** — never include direct links, contact info, or prices 6. **No superlatives** — avoid "best", "number one", "guaranteed" etc. Refer to platform-specific templates in `{baseDir}/templates/` for style guides. ## File References | File | Purpose | |------|---------| | `scripts/auth-bridge.ts` | Credential management (SocialVault required) | | `scripts/monitor.ts` | Platform monitoring orchestration | | `scripts/scorer.ts` | Multi-dimensional post scoring | | `scripts/responder.ts` | Reply generation and delivery | | `scripts/analytics.ts` | Statistics and reporting | | `scripts/adapters/*.ts` | Per-platform API/browser adapters | | `memory/brand-profile.md` | User's brand configuration | | `memory/interaction-log.jsonl` | Reply history for dedup | | `memory/blacklist.md` | Excluded users/communities/keywords | | `templates/reply-*.md` | Platform-specific reply style guides | | `references/safety-rules.md` | Rate limits and safety constraints | | `references/scoring-criteria.md` | Scoring algorithm documentation | FILE:seeddrop-v3.0.1/tsconfig.json { "compilerOptions": { "target": "ES2022", "module": "NodeNext", "moduleResolution": "NodeNext", "strict": true, "esModuleInterop": true, "skipLibCheck": true, "forceConsistentCasingInFileNames": true, "resolveJsonModule": true, "declaration": false, "outDir": "dist", "rootDir": "scripts", "noEmit": true }, "include": ["scripts/**/*.ts"], "exclude": ["node_modules", "dist"] } FILE:seeddrop-v3.0.1/templates/reply-bilibili.md # B站回复风格指南 ## 平台特征 - 年轻用户为主，二次元文化浓厚 - 弹幕文化延伸到评论区 - 喜欢有梗的表达、网络流行语 - 评论区互动活跃，气氛轻松 - 视频/专栏两种内容形态 ## 回复原则 1. **简短有力** — 100-300字，避免长篇大论 2. **有梗但不刻意** — 自然融入B站常见表达 3. **给干货** — 实用信息、具体建议 4. **真诚互动** — 像和同好聊天 5. **品牌提及极度自然** — "之前用过一个..." ## 回复模板示例 ### 模板 A — 经验分享型 ``` 干货来了！[具体建议/方法]，亲测有效。关键点在于[一个细节]，很多人容易忽略这个～ ``` ### 模板 B — 补充信息型 ``` 前排补充一下～[补充的信息/数据]，这个其实挺重要的。另外[一个小建议]也可以试试 ``` ### 模板 C — 共鸣+建议型 ``` 确实！[共鸣内容]。我的做法是[具体方案]，效果还不错。码住，有用的话记得回来点赞hh ``` ### 模板 D — 问题解答型 ``` 这个问题我刚好了解一些～[简洁的回答]。如果想深入了解的话，重点关注[方向/细节] ``` ## 禁止事项 - 不使用过于正式/商业的语气 - 不发送纯链接或引流内容 - 不使用"关注我""点赞转发"等引流话术 - 不在不相关的视频下评论 - 不重复发送相同内容 - B站评论字数限制约 233 字（普通用户） FILE:seeddrop-v3.0.1/templates/reply-tieba.md # 贴吧回复风格指南 ## 平台特征 - 按兴趣/话题分类的社区（吧） - 用户群体跨度大，从小众爱好到大众话题 - 回帖文化：直接、务实、有时略带调侃 - 每个吧有独特的氛围和潜规则 - 老牌社区，用户对广告容忍度低 ## 回复原则 1. **直接实用** — 50-200字，直击要点 2. **分条列举** — 贴吧用户偏好简洁有条理的回答 3. **匹配吧的氛围** — 不同吧说话方式不同 4. **帮忙解决问题** — 具体可操作的建议 5. **品牌提及要自然** — "我之前试过..." ## 回复模板示例 ### 模板 A — 简洁解答型 ``` 这个问题很好解决： 1. [第一步] 2. [第二步] 3. [关键注意点] 亲测有效，有问题再追问 ``` ### 模板 B — 经验分享型 ``` 我之前也遇到过一样的问题，后来发现[解决方案]效果最好。关键是[一个细节]，很多人忽略了这个。 ``` ### 模板 C — 补充型 ``` 补充一下楼上的，[额外信息]。另外建议[补充建议]，能避免不少弯路。 ``` ### 模板 D — 推荐型 ``` 这种情况建议[具体建议]，省时省力。不过要注意[注意事项]就行。 ``` ## 禁止事项 - 不使用"顶""沙发""前排"等无意义回帖 - 不放链接（会被系统删帖） - 不在不相关的帖子下回复 - 不使用"加微信""加群"等引流话术 - 不发送重复内容（系统自动删除） - 注意吧规，部分吧禁止推广 FILE:seeddrop-v3.0.1/templates/reply-xiaohongshu.md # 小红书回复风格指南 ## 平台特征 - 以女性用户为主（约 70%），年龄集中在 18-35 岁 - 生活方式导向：美妆、穿搭、美食、旅行、家居 - 视觉优先：图片/视频笔记为主 - 评论区文化：互动性强，常有追评讨论 - 语言风格：亲切、口语化、适度使用 emoji - 反感硬广：社区对"软文"和"水军"高度警觉 ## 回复原则 1. **像姐妹聊天** — 语气亲切自然，不端着 2. **给实用建议** — 具体、可操作、有细节 3. **分享真实体验** — "我之前也..."、"亲测..." 4. **适度 emoji** — 1-2 个就好，不要过度 5. **品牌提及要极其自然** — "之前试过一个..."这种模糊提法 ## 回复模板示例 ### 模板 A — 实用分享型 ``` 分享一个小方法～[具体建议]，我之前试过效果还不错！关键是要注意[细节点] 👀 ``` ### 模板 B — 经验共鸣型 ``` 哈哈我也遇到过一样的问题！后来发现[解决方案]挺好用的。你可以先从[第一步]开始试试～ ``` ### 模板 C — 补充建议型 ``` 补充一个～[额外信息/建议]。不过要注意[注意事项]，这个很多人容易踩坑 ``` ### 模板 D — 种草暗线型 ``` 这种情况建议找[品类]的专业[角色]看看，效果真的差很多。我之前帮朋友找过，[一个简短的效果描述] ``` ## 禁止事项 - 不使用"亲"、"宝贝"等淘宝客服用语 - 不放任何链接（小红书会直接折叠） - 不使用"买它"、"入手"、"安利"等明显种草词 - 不留联系方式（微信号、手机号等） - 不在笔记下刷屏式评论 - 不复制粘贴相同评论到不同笔记 - 不在美妆笔记下推非美妆服务（需要高度相关） ## 特殊注意 - 小红书评论字数限制约 500 字 - 评论被折叠后其他用户看不到（注意不要触发风控） - 新账号前 7 天评论容易被限流 - 被多人举报后可能暂时禁评 FILE:seeddrop-v3.0.1/templates/reply-zhihu.md # 知乎回复风格指南 ## 平台特征 - 知识问答社区，用户期待专业深入的内容 - 回答讲究逻辑、论据、结构化表达 - "谢邀""先说结论"等平台特有表达 - 对营销内容非常敏感，社区自净能力强 - 评论区讨论活跃，补充和反驳都很常见 ## 回复原则 1. **专业有深度** — 200-800字，结构化回答 2. **先说结论** — 开头给出核心观点 3. **提供论据** — 数据、案例、个人经验 4. **逻辑清晰** — 分段、分点、层次分明 5. **品牌提及极其隐蔽** — 仅在真正相关时自然提及 ## 回复模板示例 ### 模板 A — 专业回答型 ``` 先说结论：[核心观点]。具体来说，[详细分析/解释]。从实践角度来看： 1. [要点一] 2. [要点二] 3. [要点三] 总结一下，[一句话总结]。希望对题主有帮助。 ``` ### 模板 B — 经验分享型 ``` 正好在这个领域有些经验，分享一下。 [个人背景/经历简述] 核心建议是[建议]，原因有三： - [原因一] - [原因二] - [原因三] 最后补充一点，[补充说明]。 ``` ### 模板 C — 评论补充型 ``` 补充一个角度：[补充内容]。这一点很重要，因为[原因]。之前看到[引用来源]也提到过类似的观点。 ``` ### 模板 D — 纠正/补充型 ``` 回答整体很好，不过有一个细节需要补充：[补充内容]。实际上[更准确的信息]，这个差异在[场景]下影响比较大。 ``` ## 禁止事项 - 不使用过于口语化的表达（知乎偏书面） - 不放营销链接或自我推广 - 不使用"关注我""点赞"等引流话术 - 不发送敷衍的短回答（会被折叠） - 不在不了解的领域强行回答 - 知乎对"软文"检测严格，避免被识别为推广 FILE:seeddrop-v3.0.1/scripts/analytics.ts // SECURITY MANIFEST: // Environment variables accessed: none // External endpoints called: none // Local files read: memory/interaction-log.jsonl, memory/performance-stats.json // Local files written: memory/performance-stats.json, stdout (report) import { readFileSync, writeFileSync, existsSync, mkdirSync } from 'node:fs'; import { join, dirname } from 'node:path'; import { fileURLToPath } from 'node:url'; import type { InteractionLogEntry, PerformanceStats } from './types.js'; const __filename = fileURLToPath(import.meta.url); const __dirname = dirname(__filename); const BASE_DIR = join(__dirname, '..'); const LOG_PATH = join(BASE_DIR, 'memory', 'interaction-log.jsonl'); const STATS_PATH = join(BASE_DIR, 'memory', 'performance-stats.json'); // ─── Data Loading ─────────────────────────────────────────── function loadLog(): InteractionLogEntry[] { if (!existsSync(LOG_PATH)) return []; try { return readFileSync(LOG_PATH, 'utf-8') .split('\n') .filter((l: string) => l.trim()) .map((l: string) => JSON.parse(l) as InteractionLogEntry); } catch (err) { console.error(`[analytics] Failed to load log: ${(err as Error).message}`); return []; } } function loadStats(): PerformanceStats { if (!existsSync(STATS_PATH)) { return { total_replies: 0, by_platform: {}, by_date: {} }; } try { return JSON.parse(readFileSync(STATS_PATH, 'utf-8')) as PerformanceStats; } catch { return { total_replies: 0, by_platform: {}, by_date: {} }; } } function saveStats(stats: PerformanceStats): void { const dir = dirname(STATS_PATH); if (!existsSync(dir)) mkdirSync(dir, { recursive: true }); writeFileSync(STATS_PATH, JSON.stringify(stats, null, 2), 'utf-8'); } // ─── Report Generation ───────────────────────────────────── function filterByDateRange(entries: InteractionLogEntry[], daysBack: number): InteractionLogEntry[] { const cutoff = Date.now() - daysBack * 24 * 60 * 60 * 1000; return entries.filter(e => new Date(e.timestamp).getTime() >= cutoff); } function generateReport(entries: InteractionLogEntry[], label: string): object { const total = entries.length; const successful = entries.filter(e => e.success).length; const failed = total - successful; const byPlatform: Record = {}; const byDate: Record = {}; const byMode: Record = { approve: 0, auto: 0 }; const avgScore = total > 0 ? entries.reduce((sum, e) => sum + e.score, 0) / total : 0; for (const entry of entries) { const plat = entry.platform; if (!byPlatform[plat]) byPlatform[plat] = { total: 0, success: 0 }; byPlatform[plat].total++; if (entry.success) byPlatform[plat].success++; const dateKey = entry.timestamp.substring(0, 10); byDate[dateKey] = (byDate[dateKey] ?? 0) + 1; byMode[entry.mode] = (byMode[entry.mode] ?? 0) + 1; } return { report: label, period: { from: entries.length > 0 ? entries[0].timestamp : null, to: entries.length > 0 ? entries[entries.length - 1].timestamp : null, }, summary: { total, successful, failed, successRate: total > 0 ? Math.round((successful / total) * 100) : 0, averageScore: Math.round(avgScore * 1000) / 1000, }, byPlatform, byDate, byMode, }; } function generateTuningAdvice(entries: InteractionLogEntry[]): object { const suggestions: string[] = []; if (entries.length === 0) { return { suggestions: ['No data yet. Run some monitoring cycles first.'] }; } const avgScore = entries.reduce((s, e) => s + e.score, 0) / entries.length; if (avgScore < 0.6) { suggestions.push('Average score is low. Consider refining keywords in brand-profile.md to target more relevant discussions.'); } if (avgScore > 0.85) { suggestions.push('Average score is very high. You might lower the threshold slightly to capture more opportunities.'); } const successRate = entries.filter(e => e.success).length / entries.length; if (successRate < 0.7) { suggestions.push('Success rate is below 70%. Check credential validity and platform rate limits.'); } const platformCounts: Record = {}; for (const e of entries) { platformCounts[e.platform] = (platformCounts[e.platform] ?? 0) + 1; } const platforms = Object.keys(platformCounts); if (platforms.length === 1) { suggestions.push(`All activity is on ${platforms[0]}. Consider expanding to other platforms for broader reach.`); } const hourBuckets = new Array(24).fill(0); for (const e of entries) { const hour = new Date(e.timestamp).getHours(); hourBuckets[hour]++; } const peakHour = hourBuckets.indexOf(Math.max(...hourBuckets)); suggestions.push(`Peak activity hour: ${peakHour}:00. Consider scheduling monitoring around this time.`); if (suggestions.length === 0) { suggestions.push('Everything looks good! Keep the current configuration.'); } return { tuning: 'advice', averageScore: Math.round(avgScore * 1000) / 1000, successRate: Math.round(successRate * 100), platformDistribution: platformCounts, peakHour, suggestions, }; } // ─── Stats Update ─────────────────────────────────────────── function updateStats(entries: InteractionLogEntry[]): void { const stats = loadStats(); stats.total_replies = entries.filter(e => e.success).length; stats.by_platform = {}; stats.by_date = {}; for (const entry of entries) { if (!entry.success) continue; stats.by_platform[entry.platform] = (stats.by_platform[entry.platform] ?? 0) + 1; const dateKey = entry.timestamp.substring(0, 10); stats.by_date[dateKey] = (stats.by_date[dateKey] ?? 0) + 1; } saveStats(stats); console.error('[analytics] performance-stats.json updated'); } // ─── CLI Entry Point ──────────────────────────────────────── function main(): void { const args = process.argv.slice(2); const command = args[0] ?? 'daily'; if (command === 'test') { const log = loadLog(); const stats = loadStats(); console.log(JSON.stringify({ script: 'analytics', status: 'ok', logEntries: log.length, currentStats: stats, })); return; } const log = loadLog(); switch (command) { case 'daily': { const today = filterByDateRange(log, 1); const report = generateReport(today, 'daily'); updateStats(log); console.log(JSON.stringify(report, null, 2)); break; } case 'weekly': { const week = filterByDateRange(log, 7); const report = generateReport(week, 'weekly'); updateStats(log); console.log(JSON.stringify(report, null, 2)); break; } case 'tune': { const recent = filterByDateRange(log, 14); const advice = generateTuningAdvice(recent); console.log(JSON.stringify(advice, null, 2)); break; } default: console.error('Usage: analytics.ts '); process.exit(1); } } main(); FILE:seeddrop-v3.0.1/scripts/auth-bridge.ts // SECURITY MANIFEST: // Environment variables accessed: HOME, USERPROFILE // External endpoints called: none (SocialVault calls delegated to Agent) // Local files read: SocialVault SKILL.md (existence check only) // Local files written: none // Security: SocialVault is REQUIRED - no plaintext credential fallback import { existsSync } from 'node:fs'; import { homedir } from 'node:os'; import { join, dirname } from 'node:path'; import { fileURLToPath } from 'node:url'; import type { Credential, AuthMode, CheckResult } from './types.js'; const __filename = fileURLToPath(import.meta.url); const __dirname = dirname(__filename); const BASE_DIR = join(__dirname, '..'); // ─── SocialVault Detection ────────────────────────────────── const SOCIALVAULT_SEARCH_PATHS = [ join(homedir(), '.openclaw', 'skills', 'socialvault', 'SKILL.md'), join(homedir(), '.openclaw', 'skills', 'social-vault', 'SKILL.md'), join(homedir(), '.openclaw', 'workspace', 'skills', 'socialvault', 'SKILL.md'), join(homedir(), '.openclaw', 'workspace', 'skills', 'social-vault', 'SKILL.md'), join(BASE_DIR, '..', 'socialvault', 'SKILL.md'), join(BASE_DIR, '..', 'social-vault', 'SKILL.md'), join(BASE_DIR, '..', 'SocialVault', 'socialvault', 'SKILL.md'), ]; function detectSocialVault(): string | null { for (const p of SOCIALVAULT_SEARCH_PATHS) { if (existsSync(p)) { console.error(`[auth-bridge] SocialVault detected at: ${dirname(p)}`); return p; } } return null; } export function getAuthMode(): AuthMode { return detectSocialVault() ? 'socialvault' : 'none'; } // ─── SocialVault Mode ─────────────────────────────────────── function getSocialVaultInstruction(command: string, platform: string, profile: string): string { switch (command) { case 'use': return `socialvault use ${platform}-${profile}`; case 'token': return `socialvault token ${platform}-${profile}`; case 'release': return `socialvault release ${platform}-${profile}`; case 'check': return `socialvault check ${platform}-${profile}`; default: return `socialvault status`; } } // ─── Public API ───────────────────────────────────────────── export function getCredential(platform: string, profile: string = 'default'): Credential | null { const mode = getAuthMode(); if (mode === 'socialvault') { const instruction = getSocialVaultInstruction('token', platform, profile); console.error(`[auth-bridge] SocialVault mode — Agent should run: ${instruction}`); return { authType: 'oauth', value: `__socialvault_pending__:${instruction}`, profile, source: 'socialvault', }; } console.error(`[auth-bridge] SocialVault is required but not detected. Please install: clawhub install socialvault`); return null; } export function checkCredential(platform: string, profile: string = 'default'): CheckResult { const cred = getCredential(platform, profile); if (!cred) { return { valid: false, error: `No credential found for ${platform}/${profile}` }; } if (cred.source === 'socialvault') { return { valid: true, username: `(via SocialVault, run: socialvault check ${platform}-${profile})` }; } return { valid: cred.value.length > 0, error: cred.value.length === 0 ? 'Credential value is empty' : undefined, }; } // ─── CLI Entry Point ──────────────────────────────────────── const IS_MAIN = process.argv[1]?.replace(/\\/g, '/').endsWith('auth-bridge.ts'); function main(): void { if (!IS_MAIN) return; const args = process.argv.slice(2); const command = args[0]; switch (command) { case 'mode': console.log(JSON.stringify({ mode: getAuthMode() })); break; case 'get': { const platform = args[1]; const profile = args[2] ?? 'default'; if (!platform) { console.error('Usage: auth-bridge.ts get [profile]'); process.exit(1); } const cred = getCredential(platform, profile); if (!cred) { console.error(`[auth-bridge] Failed to get credential for ${platform}/${profile}`); process.exit(1); } console.log(JSON.stringify(cred)); break; } case 'check': { const platform = args[1]; const profile = args[2] ?? 'default'; if (!platform) { console.error('Usage: auth-bridge.ts check [profile]'); process.exit(1); } const result = checkCredential(platform, profile); console.log(JSON.stringify(result)); break; } case 'test': console.log(JSON.stringify({ script: 'auth-bridge', status: 'ok', mode: getAuthMode(), socialVaultDetected: detectSocialVault() !== null, })); break; default: console.error('Usage: auth-bridge.ts [args]'); console.error(' mode — Show auth mode (local/socialvault)'); console.error(' get [profile] — Get credential'); console.error(' check [profile] — Check credential validity'); console.error(' test — Self-test'); process.exit(1); } } main(); FILE:seeddrop-v3.0.1/scripts/monitor.ts // SECURITY MANIFEST: // Environment variables accessed: none // External endpoints called: none (delegated to adapters) // Local files read: memory/brand-profile.md, memory/interaction-log.jsonl, memory/blacklist.md // Local files written: stdout (JSONL) import { readFileSync, existsSync } from 'node:fs'; import { join, dirname } from 'node:path'; import { fileURLToPath } from 'node:url'; import { getAdapter, listPlatforms } from './adapters/base.js'; import { getCredential } from './auth-bridge.js'; import type { Post, InteractionLogEntry } from './types.js'; const __filename = fileURLToPath(import.meta.url); const __dirname = dirname(__filename); const BASE_DIR = join(__dirname, '..'); // ─── Brand Profile ────────────────────────────────────────── interface MonitorConfig { keywords: string[]; platforms: string[]; timeRange: string; } function loadConfig(): MonitorConfig { const profilePath = join(BASE_DIR, 'memory', 'brand-profile.md'); const defaults: MonitorConfig = { keywords: [], platforms: [], timeRange: 'day', }; if (!existsSync(profilePath)) { console.error('[monitor] brand-profile.md not found'); return defaults; } try { const raw = readFileSync(profilePath, 'utf-8'); const kwMatch = raw.match(/##\s*关键词.*?\n([\s\S]*?)(?=\n##|\n$|$)/i) ?? raw.match(/##\s*Keywords.*?\n([\s\S]*?)(?=\n##|\n$|$)/i); if (kwMatch) { defaults.keywords = kwMatch[1] .split('\n') .map((l: string) => l.replace(/^[-*]\s*/, '').trim()) .filter((l: string) => l.length > 0); } const platMatch = raw.match(/##\s*平台.*?\n([\s\S]*?)(?=\n##|\n$|$)/i) ?? raw.match(/##\s*Platforms.*?\n([\s\S]*?)(?=\n##|\n$|$)/i); if (platMatch) { defaults.platforms = platMatch[1] .split('\n') .map((l: string) => l.replace(/^[-*]\s*/, '').trim()) .filter((l: string) => l.length > 0); } return defaults; } catch (err) { console.error(`[monitor] Failed to load config: ${(err as Error).message}`); return defaults; } } // ─── Deduplication ────────────────────────────────────────── function loadRepliedPostIds(): Set { const logPath = join(BASE_DIR, 'memory', 'interaction-log.jsonl'); if (!existsSync(logPath)) return new Set(); try { const raw = readFileSync(logPath, 'utf-8'); const ids = new Set(); for (const line of raw.split('\n')) { if (!line.trim()) continue; try { const entry = JSON.parse(line) as InteractionLogEntry; ids.add(entry.postId); } catch { /* skip malformed lines */ } } return ids; } catch { return new Set(); } } // ─── Blacklist ────────────────────────────────────────────── interface Blacklist { users: string[]; communities: string[]; keywords: string[]; } function loadBlacklist(): Blacklist { const path = join(BASE_DIR, 'memory', 'blacklist.md'); const defaults: Blacklist = { users: [], communities: [], keywords: [] }; if (!existsSync(path)) return defaults; try { const raw = readFileSync(path, 'utf-8'); let currentSection = ''; for (const line of raw.split('\n')) { const sectionMatch = line.match(/^##\s*(.+)/); if (sectionMatch) { currentSection = sectionMatch[1].toLowerCase(); continue; } const item = line.replace(/^[-*]\s*/, '').trim(); if (!item || item.startsWith('#')) continue; if (currentSection.includes('用户') || currentSection.includes('user')) { defaults.users.push(item.toLowerCase()); } else if (currentSection.includes('社区') || currentSection.includes('communit') || currentSection.includes('subreddit')) { defaults.communities.push(item.toLowerCase()); } else if (currentSection.includes('关键词') || currentSection.includes('keyword')) { defaults.keywords.push(item.toLowerCase()); } } return defaults; } catch { return defaults; } } function isBlacklisted(post: Post, blacklist: Blacklist): boolean { if (blacklist.users.includes(post.author.toLowerCase())) return true; if (post.community && blacklist.communities.includes(post.community.toLowerCase())) return true; const text = `${post.title} ${post.body}`.toLowerCase(); for (const kw of blacklist.keywords) { if (text.includes(kw)) return true; } return false; } // ─── Main Monitor Logic ───────────────────────────────────── async function monitorPlatform( platformId: string, keywords: string[], timeRange: string, target?: string, ): Promise { const adapter = await getAdapter(platformId); const cred = getCredential(platformId); if (!cred) { console.error(`[monitor] No credentials for ${platformId}, skipping`); return []; } if (cred.value.startsWith('__socialvault_pending__:')) { console.error(`[monitor] SocialVault credential pending for ${platformId}`); console.error(`[monitor] Agent should run: ${cred.value.replace('__socialvault_pending__:', '')}`); console.error(`[monitor] Then re-run monitor with valid credentials`); return []; } const allPosts: Post[] = []; for (const keyword of keywords) { console.error(`[monitor] Searching ${platformId} for "${keyword}" (range: ${timeRange})${target ? ` in ${target}` : ''}`); const posts = await adapter.search(keyword, timeRange, cred, target); allPosts.push(...posts); } const uniquePosts = new Map(); for (const post of allPosts) { if (!uniquePosts.has(post.id)) { uniquePosts.set(post.id, post); } } return Array.from(uniquePosts.values()); } // ─── CLI Entry Point ──────────────────────────────────────── async function main(): Promise { const args = process.argv.slice(2); if (args[0] === 'test') { const config = loadConfig(); const blacklist = loadBlacklist(); const repliedIds = loadRepliedPostIds(); console.log(JSON.stringify({ script: 'monitor', status: 'ok', config, blacklistCounts: { users: blacklist.users.length, communities: blacklist.communities.length, keywords: blacklist.keywords.length, }, repliedPostCount: repliedIds.size, availablePlatforms: listPlatforms(), })); return; } const platformArg = args[0]; const targetArg = args[1]; // e.g. specific community (吧名, 分区等) if (!platformArg) { console.error('Usage: monitor.ts [target]'); console.error(`Available platforms: ${listPlatforms().join(', ')}`); process.exit(1); } const config = loadConfig(); if (config.keywords.length === 0) { console.error('[monitor] No keywords configured in brand-profile.md'); console.error('[monitor] Run "seeddrop setup" to configure keywords'); process.exit(1); } const blacklist = loadBlacklist(); const repliedIds = loadRepliedPostIds(); const platforms = platformArg === 'all' ? (config.platforms.length > 0 ? config.platforms : listPlatforms()) : [platformArg]; let totalOutput = 0; for (const plat of platforms) { try { const posts = await monitorPlatform(plat, config.keywords, config.timeRange, targetArg); console.error(`[monitor] ${plat}: found ${posts.length} raw posts`); for (const post of posts) { if (repliedIds.has(post.id)) { console.error(`[monitor] SKIP ${post.id}: already replied`); continue; } if (isBlacklisted(post, blacklist)) { console.error(`[monitor] SKIP ${post.id}: blacklisted`); continue; } console.log(JSON.stringify(post)); totalOutput++; } } catch (err) { console.error(`[monitor] Error monitoring ${plat}: ${(err as Error).message}`); } } console.error(`[monitor] Total output: ${totalOutput} posts`); } main().catch(err => { console.error(`[monitor] Fatal error: ${(err as Error).message}`); process.exit(1); }); FILE:seeddrop-v3.0.1/scripts/responder.ts // SECURITY MANIFEST: // Environment variables accessed: none // External endpoints called: none (reply sending delegated to adapters via Agent) // Local files read: stdin (JSONL), memory/interaction-log.jsonl, memory/brand-profile.md // Local files written: stdout (drafts/results), memory/interaction-log.jsonl import { readFileSync, appendFileSync, existsSync, mkdirSync } from 'node:fs'; import { join, dirname } from 'node:path'; import { fileURLToPath } from 'node:url'; import { createInterface } from 'node:readline'; import type { ScoredPost, InteractionLogEntry, ReplyDraft, ReplyMode, } from './types.js'; import { PLATFORM_DAILY_LIMITS } from './types.js'; const __filename = fileURLToPath(import.meta.url); const __dirname = dirname(__filename); const BASE_DIR = join(__dirname, '..'); const LOG_PATH = join(BASE_DIR, 'memory', 'interaction-log.jsonl'); // ─── Interaction Log ──────────────────────────────────────── function loadInteractionLog(): InteractionLogEntry[] { if (!existsSync(LOG_PATH)) return []; try { const raw = readFileSync(LOG_PATH, 'utf-8'); return raw .split('\n') .filter((l: string) => l.trim()) .map((l: string) => JSON.parse(l) as InteractionLogEntry); } catch (err) { console.error(`[responder] Failed to load interaction log: ${(err as Error).message}`); return []; } } export function appendToLog(entry: InteractionLogEntry): void { const dir = dirname(LOG_PATH); if (!existsSync(dir)) mkdirSync(dir, { recursive: true }); appendFileSync(LOG_PATH, JSON.stringify(entry) + '\n', 'utf-8'); } // ─── Deduplication & Safety Checks ────────────────────────── function isAlreadyReplied(postId: string, log: InteractionLogEntry[]): boolean { return log.some(entry => entry.postId === postId); } function isAuthorCoolingDown(author: string, platform: string, log: InteractionLogEntry[]): boolean { const cutoff = Date.now() - 24 * 60 * 60 * 1000; return log.some( entry => entry.author === author && entry.platform === platform && new Date(entry.timestamp).getTime() > cutoff, ); } function getTodayReplyCount(platform: string, log: InteractionLogEntry[]): number { const todayStart = new Date(); todayStart.setHours(0, 0, 0, 0); const cutoff = todayStart.getTime(); return log.filter( entry => entry.platform === platform && entry.success && new Date(entry.timestamp).getTime() >= cutoff, ).length; } function getDailyLimit(platform: string, mode: ReplyMode): number { const limits = PLATFORM_DAILY_LIMITS[platform] ?? PLATFORM_DAILY_LIMITS['_default']!; return mode === 'auto' ? limits.auto : limits.approve; } // ─── Reply Draft Generation ───────────────────────────────── function generateDraft(post: ScoredPost): ReplyDraft { return { postId: post.id, postUrl: post.url, postTitle: post.title, platform: post.platform, content: `__DRAFT_PLACEHOLDER__`, score: post.finalScore, }; } // ─── CLI Entry Point ──────────────────────────────────────── async function main(): Promise { const args = process.argv.slice(2); if (args[0] === 'test') { const log = loadInteractionLog(); console.log(JSON.stringify({ script: 'responder', status: 'ok', logEntries: log.length, logPath: LOG_PATH, })); return; } // SECURITY: Only approve mode is supported - no auto-reply to prevent abuse const mode: ReplyMode = 'approve'; if (args[0] === 'auto') { console.error('[responder] SECURITY: Auto mode is disabled. All replies require manual approval.'); process.exit(1); } console.error(`[responder] Mode: ${mode} (manual approval required)`); const log = loadInteractionLog(); const rl = createInterface({ input: process.stdin, terminal: false }); let inputCount = 0; let draftCount = 0; let skipDup = 0; let skipAuthor = 0; let skipLimit = 0; for await (const line of rl) { if (!line.trim()) continue; try { const post = JSON.parse(line) as ScoredPost; inputCount++; if (isAlreadyReplied(post.id, log)) { console.error(`[responder] SKIP ${post.id}: already replied`); skipDup++; continue; } if (isAuthorCoolingDown(post.author, post.platform, log)) { console.error(`[responder] SKIP ${post.id}: author "${post.author}" cooldown (24h)`); skipAuthor++; continue; } const dailyLimit = getDailyLimit(post.platform, mode); const todayCount = getTodayReplyCount(post.platform, log); if (todayCount >= dailyLimit) { console.error(`[responder] SKIP ${post.id}: daily limit reached (${todayCount}/${dailyLimit})`); skipLimit++; continue; } const draft = generateDraft(post); if (mode === 'approve') { console.log(JSON.stringify({ action: 'review_draft', draft, post: { id: post.id, url: post.url, title: post.title, body: post.body.substring(0, 300), author: post.author, platform: post.platform, community: post.community, scores: post.scores, finalScore: post.finalScore, }, instructions: `Generate a value-first reply for this ${post.platform} post. Follow the template at templates/reply-${post.platform}.md. Brand mention ≤20%. Then present the draft to the user for approval.`, })); } else { console.log(JSON.stringify({ action: 'auto_reply', draft, post: { id: post.id, url: post.url, title: post.title, body: post.body.substring(0, 300), author: post.author, platform: post.platform, community: post.community, scores: post.scores, finalScore: post.finalScore, }, instructions: `Generate and send a value-first reply for this ${post.platform} post. Follow the template at templates/reply-${post.platform}.md. Brand mention ≤20%. Log the result to interaction-log.jsonl.`, })); } draftCount++; } catch { console.error(`[responder] Failed to parse line: ${line.substring(0, 80)}`); } } console.error(`[responder] Done: ${draftCount} drafts from ${inputCount} posts (skipped: ${skipDup} dup, ${skipAuthor} author, ${skipLimit} limit)`); } main().catch(err => { console.error(`[responder] Fatal error: ${(err as Error).message}`); process.exit(1); }); FILE:seeddrop-v3.0.1/scripts/scorer.ts // SECURITY MANIFEST: // Environment variables accessed: none // External endpoints called: none // Local files read: stdin (JSONL), memory/brand-profile.md // Local files written: stdout (JSONL) import { readFileSync, existsSync } from 'node:fs'; import { join, dirname } from 'node:path'; import { fileURLToPath } from 'node:url'; import { createInterface } from 'node:readline'; import type { Post, ScoredPost, ScoreBreakdown } from './types.js'; import { SCORE_WEIGHTS, DEFAULT_THRESHOLD, AUTO_MODE_MIN_THRESHOLD, AUTO_MODE_MIN_RISK, } from './types.js'; const __filename = fileURLToPath(import.meta.url); const __dirname = dirname(__filename); const BASE_DIR = join(__dirname, '..'); // ─── Brand Profile Loading ────────────────────────────────── interface ParsedBrandProfile { keywords: string[]; mode: 'approve' | 'auto'; threshold: number; } function loadBrandProfile(): ParsedBrandProfile { const profilePath = join(BASE_DIR, 'memory', 'brand-profile.md'); const defaults: ParsedBrandProfile = { keywords: [], mode: 'approve', threshold: DEFAULT_THRESHOLD, }; if (!existsSync(profilePath)) { console.error('[scorer] brand-profile.md not found, using defaults'); return defaults; } try { const raw = readFileSync(profilePath, 'utf-8'); const keywordsMatch = raw.match(/##\s*关键词.*?\n([\s\S]*?)(?=\n##|\n$|$)/i) ?? raw.match(/##\s*Keywords.*?\n([\s\S]*?)(?=\n##|\n$|$)/i); if (keywordsMatch) { defaults.keywords = keywordsMatch[1] .split('\n') .map((l: string) => l.replace(/^[-*]\s*/, '').trim()) .filter((l: string) => l.length > 0); } const modeMatch = raw.match(/模式[：:]\s*(approve|auto)/i) ?? raw.match(/mode[：:]\s*(approve|auto)/i); if (modeMatch) { defaults.mode = modeMatch[1].toLowerCase() as 'approve' | 'auto'; } const thresholdMatch = raw.match(/阈值[：:]\s*([\d.]+)/i) ?? raw.match(/threshold[：:]\s*([\d.]+)/i); if (thresholdMatch) { defaults.threshold = parseFloat(thresholdMatch[1]); } return defaults; } catch (err) { console.error(`[scorer] Failed to load brand profile: ${(err as Error).message}`); return defaults; } } // ─── Scoring Functions ────────────────────────────────────── const INTENT_KEYWORDS = { high: ['求推荐', '怎么', '有什么好的', '推荐一下', '求助', '请问', '有没有推荐', '怎么选', '怎么办', 'how to', 'recommend', 'help', 'looking for'], medium: ['讨论', '大家觉得', '有没有人用过', '有经验', '分享一下', '什么体验', '值不值', 'what do you think', 'experience'], low: ['坑', '吐槽', '垃圾', '难用', '太差了', '后悔', '避雷', 'frustrated', 'terrible'], }; function scoreRelevance(post: Post, keywords: string[]): number { if (keywords.length === 0) return 0.5; const text = `${post.title} ${post.body}`.toLowerCase(); let hits = 0; for (const kw of keywords) { if (text.includes(kw.toLowerCase())) hits++; } const ratio = hits / keywords.length; if (ratio > 0.5) return 1.0; if (ratio > 0.2) return 0.7; if (ratio > 0) return 0.5; return 0.2; } function scoreIntent(post: Post): number { const text = `${post.title} ${post.body}`.toLowerCase(); for (const kw of INTENT_KEYWORDS.high) { if (text.includes(kw)) return 0.9; } for (const kw of INTENT_KEYWORDS.medium) { if (text.includes(kw)) return 0.7; } for (const kw of INTENT_KEYWORDS.low) { if (text.includes(kw)) return 0.5; } return 0.3; } function scoreFreshness(post: Post): number { const ageMs = Date.now() - new Date(post.createdAt).getTime(); const ageHours = ageMs / (1000 * 60 * 60); if (ageHours <= 2) return 1.0; if (ageHours <= 6) return 0.9; if (ageHours <= 12) return 0.8; if (ageHours <= 24) return 0.6; if (ageHours <= 48) return 0.4; return 0.2; } function scoreRisk(post: Post): number { const text = `${post.title} ${post.body}`.toLowerCase(); const officialMarkers = ['[meta]', '[announcement]', '[mod]', '[official]', '置顶', 'pinned']; for (const marker of officialMarkers) { if (text.includes(marker)) return 0.3; } const controversialMarkers = ['politic', 'religion', 'controversial', '政治', '宗教', '敏感']; for (const marker of controversialMarkers) { if (text.includes(marker)) return 0.2; } return 0.9; } export function scorePost(post: Post, keywords: string[]): ScoredPost { const scores: ScoreBreakdown = { relevance: scoreRelevance(post, keywords), intent: scoreIntent(post), freshness: scoreFreshness(post), risk: scoreRisk(post), }; const finalScore = scores.relevance * SCORE_WEIGHTS.relevance + scores.intent * SCORE_WEIGHTS.intent + scores.freshness * SCORE_WEIGHTS.freshness + scores.risk * SCORE_WEIGHTS.risk; return { ...post, scores, finalScore: Math.round(finalScore * 1000) / 1000 }; } // ─── CLI Entry Point ──────────────────────────────────────── async function main(): Promise { const args = process.argv.slice(2); if (args[0] === 'test') { const testPost: Post = { id: 'test1', url: 'https://reddit.com/r/test/test1', title: 'How to build a landing page?', body: 'Looking for recommendations on tools to build a quick landing page.', author: 'testuser', createdAt: new Date().toISOString(), platform: 'reddit', }; const result = scorePost(testPost, ['landing page', 'build', 'tool']); console.log(JSON.stringify({ script: 'scorer', status: 'ok', testResult: result, })); return; } const thresholdArg = parseFloat(args[0] || ''); const profile = loadBrandProfile(); const threshold = !isNaN(thresholdArg) ? thresholdArg : profile.threshold; const effectiveThreshold = profile.mode === 'auto' ? Math.max(threshold, AUTO_MODE_MIN_THRESHOLD) : threshold; console.error(`[scorer] Mode: ${profile.mode}, Threshold: ${effectiveThreshold}, Keywords: ${profile.keywords.length}`); const rl = createInterface({ input: process.stdin, terminal: false }); let inputCount = 0; let passedCount = 0; for await (const line of rl) { if (!line.trim()) continue; try { const post = JSON.parse(line) as Post; inputCount++; const scored = scorePost(post, profile.keywords); if (scored.finalScore < effectiveThreshold) { console.error(`[scorer] SKIP ${scored.id} (score=${scored.finalScore} < ${effectiveThreshold})`); continue; } if (profile.mode === 'auto' && scored.scores.risk < AUTO_MODE_MIN_RISK) { console.error(`[scorer] SKIP ${scored.id} (risk=${scored.scores.risk} < ${AUTO_MODE_MIN_RISK}, auto mode)`); continue; } passedCount++; console.log(JSON.stringify(scored)); } catch { console.error(`[scorer] Failed to parse line: ${line.substring(0, 80)}`); } } console.error(`[scorer] Done: ${passedCount}/${inputCount} posts passed threshold`); } main().catch(err => { console.error(`[scorer] Fatal error: ${(err as Error).message}`); process.exit(1); }); FILE:seeddrop-v3.0.1/scripts/types.ts // SECURITY MANIFEST: // Environment variables accessed: none // External endpoints called: none // Local files read: none // Local files written: none // ─── Credential ───────────────────────────────────────────── export interface Credential { authType: 'api_token' | 'cookie' | 'oauth'; value: string; profile?: string; source: 'socialvault' | 'local'; } // ─── Post & Scoring ───────────────────────────────────────── export interface Post { id: string; url: string; title: string; body: string; author: string; createdAt: string; // ISO 8601 platform: string; community?: string; // 吧名、知乎话题、B站分区等 metadata?: Record; } export interface ScoreBreakdown { relevance: number; intent: number; freshness: number; risk: number; } export interface ScoredPost extends Post { scores: ScoreBreakdown; finalScore: number; } // ─── Reply ────────────────────────────────────────────────── export interface ReplyResult { success: boolean; replyId?: string; error?: string; mode?: 'api' | 'browser'; } export interface ReplyDraft { postId: string; postUrl: string; postTitle: string; platform: string; content: string; score: number; } // ─── Auth / Check ─────────────────────────────────────────── export interface CheckResult { valid: boolean; username?: string; error?: string; } export type AuthMode = 'socialvault' | 'none'; // ─── Rate Limiting ────────────────────────────────────────── export interface RateLimitInfo { requestsPerMinute: number; repliesPerDay: number; minReplyIntervalSeconds: number; notes: string; } export interface DailyLimits { approve: number; auto: number; } export const PLATFORM_DAILY_LIMITS: Record = { 'bilibili': { approve: 30, auto: 15 }, 'tieba': { approve: 20, auto: 10 }, 'zhihu': { approve: 10, auto: 5 }, 'xiaohongshu': { approve: 10, auto: 5 }, '_default': { approve: 10, auto: 5 }, }; // ─── Platform Adapter ─────────────────────────────────────── export interface PlatformAdapter { readonly platformId: string; readonly platformName: string; search( keyword: string, timeRange: string, credential: Credential, target?: string, ): Promise; reply( postId: string, content: string, credential: Credential, ): Promise; check(credential: Credential): Promise; rateLimitInfo(): RateLimitInfo; } // ─── Interaction Log ──────────────────────────────────────── export interface InteractionLogEntry { timestamp: string; // ISO 8601 platform: string; postId: string; postUrl: string; postTitle: string; author: string; replyContent: string; replyId?: string; score: number; mode: 'approve' | 'auto'; success: boolean; } // ─── Brand Profile ────────────────────────────────────────── export interface BrandProfile { businessName: string; description: string; keywords: string[]; platforms: string[]; mode: 'approve' | 'auto'; threshold: number; language: string; } // ─── Scoring Config ───────────────────────────────────────── export const SCORE_WEIGHTS = { relevance: 0.35, intent: 0.30, freshness: 0.20, risk: 0.15, } as const; export const DEFAULT_THRESHOLD = 0.6; export const AUTO_MODE_MIN_THRESHOLD = 0.7; export const AUTO_MODE_MIN_RISK = 0.5; // ─── Browser Instruction (for adapter browser mode) ───────── export const BROWSER_INSTRUCTION_ID = '__browser_instruction__'; export interface BrowserStep { action: 'navigate' | 'wait' | 'extract' | 'click' | 'type'; url?: string; selector?: string; fields?: string[]; text?: string; } export interface BrowserInstruction { mode: 'browser'; action: 'search' | 'reply' | 'check'; steps: BrowserStep[]; cookies?: string; } // ─── Cookie Parsing ───────────────────────────────────────── export function parseCookieValue(raw: string, name: string): string | undefined { const match = raw.match(new RegExp(`(?:^|;\\s*)${name}=([^;]*)`)); return match?.[1]; } // ─── Utility types ────────────────────────────────────────── export type ReplyMode = 'approve' | 'auto'; export interface PerformanceStats { total_replies: number; by_platform: Record; by_date: Record; } FILE:seeddrop-v3.0.1/scripts/adapters/base.ts // SECURITY MANIFEST: // Environment variables accessed: none // External endpoints called: none // Local files read: none // Local files written: none import type { PlatformAdapter } from '../types.js'; const adapterRegistry: Record Promise> = { 'bilibili': async () => { const { BilibiliAdapter } = await import('./bilibili.js'); return new BilibiliAdapter(); }, 'tieba': async () => { const { TiebaAdapter } = await import('./tieba.js'); return new TiebaAdapter(); }, 'zhihu': async () => { const { ZhihuAdapter } = await import('./zhihu.js'); return new ZhihuAdapter(); }, 'xiaohongshu': async () => { const { XiaohongshuAdapter } = await import('./xiaohongshu.js'); return new XiaohongshuAdapter(); }, }; export async function getAdapter(platformId: string): Promise { const factory = adapterRegistry[platformId]; if (!factory) { const available = Object.keys(adapterRegistry).join(', '); throw new Error(`Unknown platform: "${platformId}". Available: ${available}`); } return factory(); } export function listPlatforms(): string[] { return Object.keys(adapterRegistry); } const isMainModule = process.argv[1]?.replace(/\\/g, '/').endsWith('adapters/base.ts'); if (isMainModule && process.argv[2] === 'test') { console.log(JSON.stringify({ script: 'adapters/base', status: 'ok', platforms: listPlatforms(), })); } FILE:seeddrop-v3.0.1/scripts/adapters/bilibili.ts // SECURITY MANIFEST: // Environment variables accessed: none // External endpoints called: https://api.bilibili.com // Local files read: none // Local files written: none import type { PlatformAdapter, Credential, Post, ReplyResult, CheckResult, RateLimitInfo, } from '../types.js'; import { parseCookieValue } from '../types.js'; const BILIBILI_API = 'https://api.bilibili.com'; function buildHeaders(credential: Credential): Record { return { 'Cookie': credential.value, 'User-Agent': 'Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/120.0.0.0 Safari/537.36', 'Referer': 'https://www.bilibili.com', }; } export class BilibiliAdapter implements PlatformAdapter { readonly platformId = 'bilibili'; readonly platformName = '哔哩哔哩'; async search( keyword: string, _timeRange: string, credential: Credential, _target?: string, ): Promise { try { const params = new URLSearchParams({ keyword, search_type: 'video', page: '1', page_size: '20', }); const url = `${BILIBILI_API}/x/web-interface/wbi/search/type?${params}`; const response = await fetch(url, { headers: buildHeaders(credential) }); if (!response.ok) { console.error(`[bilibili] Search failed: ${response.status}`); return []; } const data = await response.json() as { code: number; data?: { result?: Array<{ aid: number; bvid: string; title: string; description: string; author: string; pubdate: number; typeid: number; typename: string; }>; }; }; if (data.code !== 0 || !data.data?.result) { console.error(`[bilibili] Search API error: code=${data.code}`); return []; } return data.data.result.map(item => ({ id: String(item.aid), url: `https://www.bilibili.com/video/${item.bvid}`, title: item.title.replace(/<[^>]*>/g, ''), body: item.description, author: item.author, createdAt: new Date(item.pubdate * 1000).toISOString(), platform: this.platformId, community: item.typename, metadata: { bvid: item.bvid, aid: item.aid }, })); } catch (error) { console.error(`[bilibili] Search error: ${(error as Error).message}`); return []; } } async reply( postId: string, content: string, credential: Credential, ): Promise { try { const csrf = parseCookieValue(credential.value, 'bili_jct'); if (!csrf) { return { success: false, error: 'Missing bili_jct in cookie (required for CSRF)', mode: 'api' }; } const response = await fetch(`${BILIBILI_API}/x/v2/reply/add`, { method: 'POST', headers: { ...buildHeaders(credential), 'Content-Type': 'application/x-www-form-urlencoded', }, body: new URLSearchParams({ type: '1', oid: postId, message: content, csrf, }).toString(), }); if (!response.ok) { return { success: false, error: `HTTP ${response.status}`, mode: 'api' }; } const data = await response.json() as { code: number; message?: string; data?: { rpid?: number }; }; if (data.code !== 0) { return { success: false, error: `B站 API: ${data.message ?? `code ${data.code}`}`, mode: 'api' }; } return { success: true, replyId: data.data?.rpid ? String(data.data.rpid) : undefined, mode: 'api', }; } catch (error) { return { success: false, error: (error as Error).message, mode: 'api' }; } } async check(credential: Credential): Promise { try { const response = await fetch(`${BILIBILI_API}/x/web-interface/nav`, { headers: buildHeaders(credential), }); if (!response.ok) return { valid: false, error: `HTTP ${response.status}` }; const data = await response.json() as { code: number; data?: { isLogin: boolean; uname?: string }; }; if (data.code !== 0 || !data.data?.isLogin) { return { valid: false, error: 'Not logged in or cookie expired' }; } return { valid: true, username: data.data.uname }; } catch (error) { return { valid: false, error: (error as Error).message }; } } rateLimitInfo(): RateLimitInfo { return { requestsPerMinute: 30, repliesPerDay: 30, minReplyIntervalSeconds: 60, notes: 'B站: 搜索≤30次/时, 评论≤50条/天, 重复内容自动过滤', }; } } const isMainModule = process.argv[1]?.replace(/\\/g, '/').endsWith('bilibili.ts'); if (isMainModule && process.argv[2] === 'test') { const adapter = new BilibiliAdapter(); console.log(JSON.stringify({ adapter: 'bilibili', status: 'ok', platformId: adapter.platformId, platformName: adapter.platformName, rateLimit: adapter.rateLimitInfo(), })); } FILE:seeddrop-v3.0.1/scripts/adapters/tieba.ts // SECURITY MANIFEST: // Environment variables accessed: none // External endpoints called: https://tieba.baidu.com // Local files read: none // Local files written: none import type { PlatformAdapter, Credential, Post, ReplyResult, CheckResult, RateLimitInfo, } from '../types.js'; const TIEBA_BASE = 'https://tieba.baidu.com'; function buildHeaders(credential: Credential): Record { return { 'Cookie': credential.value, 'User-Agent': 'Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/120.0.0.0 Safari/537.36', }; } async function getTbs(credential: Credential): Promise { try { const response = await fetch(`${TIEBA_BASE}/dc/common/tbs`, { headers: buildHeaders(credential), }); if (!response.ok) return null; const data = await response.json() as { tbs?: string; is_login?: number }; return data.tbs ?? null; } catch { return null; } } export class TiebaAdapter implements PlatformAdapter { readonly platformId = 'tieba'; readonly platformName = '百度贴吧'; async search( keyword: string, _timeRange: string, credential: Credential, target?: string, ): Promise { try { const url = target ? `${TIEBA_BASE}/f?kw=${encodeURIComponent(target)}&ie=utf-8&pn=0` : `${TIEBA_BASE}/f/search/res?qw=${encodeURIComponent(keyword)}&rn=20&pn=1`; const response = await fetch(url, { headers: buildHeaders(credential) }); if (!response.ok) { console.error(`[tieba] Search failed: ${response.status}`); return []; } const html = await response.text(); const posts: Post[] = []; const threadPattern = /href="\/p\/(\d+)"[^>]*>([^<]+)<\/a>/g; let match; while ((match = threadPattern.exec(html)) !== null) { const [, threadId, title] = match; if (!threadId || !title) continue; const cleanTitle = title.trim(); if (cleanTitle.length < 4) continue; posts.push({ id: threadId, url: `${TIEBA_BASE}/p/${threadId}`, title: cleanTitle, body: '', author: '', createdAt: new Date().toISOString(), platform: this.platformId, community: target, }); } const uniquePosts = new Map(); for (const p of posts) { if (!uniquePosts.has(p.id)) uniquePosts.set(p.id, p); } return Array.from(uniquePosts.values()).slice(0, 20); } catch (error) { console.error(`[tieba] Search error: ${(error as Error).message}`); return []; } } async reply( postId: string, content: string, credential: Credential, ): Promise { try { const tbs = await getTbs(credential); if (!tbs) { return { success: false, error: 'Failed to get tbs token', mode: 'api' }; } const metadata = (this as unknown as { _lastReplyMeta?: { kw: string; fid: string } })._lastReplyMeta; const kw = metadata?.kw ?? ''; const fid = metadata?.fid ?? ''; const response = await fetch(`${TIEBA_BASE}/f/commit/post/add`, { method: 'POST', headers: { ...buildHeaders(credential), 'Content-Type': 'application/x-www-form-urlencoded', }, body: new URLSearchParams({ ie: 'utf-8', kw, fid, tid: postId, content, tbs, }).toString(), }); if (!response.ok) { return { success: false, error: `HTTP ${response.status}`, mode: 'api' }; } const data = await response.json() as { err_code?: number; error?: string; data?: { tid?: string }; }; if (data.err_code && data.err_code !== 0) { return { success: false, error: data.error ?? `err_code ${data.err_code}`, mode: 'api' }; } return { success: true, mode: 'api' }; } catch (error) { return { success: false, error: (error as Error).message, mode: 'api' }; } } async check(credential: Credential): Promise { try { const response = await fetch(`${TIEBA_BASE}/dc/common/tbs`, { headers: buildHeaders(credential), }); if (!response.ok) return { valid: false, error: `HTTP ${response.status}` }; const data = await response.json() as { tbs?: string; is_login?: number }; if (data.is_login !== 1) { return { valid: false, error: 'Not logged in or BDUSS expired' }; } return { valid: true, username: '(logged in via BDUSS)' }; } catch (error) { return { valid: false, error: (error as Error).message }; } } rateLimitInfo(): RateLimitInfo { return { requestsPerMinute: 20, repliesPerDay: 20, minReplyIntervalSeconds: 120, notes: '贴吧: BDUSS有效期6个月+, 回帖≤30条/天, 重复内容被删, 需从请求头获取BDUSS(非Cookie-Editor的BDUSS_BFESS)', }; } } const isMainModule = process.argv[1]?.replace(/\\/g, '/').endsWith('tieba.ts'); if (isMainModule && process.argv[2] === 'test') { const adapter = new TiebaAdapter(); console.log(JSON.stringify({ adapter: 'tieba', status: 'ok', platformId: adapter.platformId, platformName: adapter.platformName, rateLimit: adapter.rateLimitInfo(), })); } FILE:seeddrop-v3.0.1/scripts/adapters/xiaohongshu.ts // SECURITY MANIFEST: // Environment variables accessed: none // External endpoints called: https://edith.xiaohongshu.com/api/sns/web/v1 // Local files read: none // Local files written: none import type { PlatformAdapter, Credential, Post, ReplyResult, CheckResult, RateLimitInfo, BrowserInstruction, } from '../types.js'; import { BROWSER_INSTRUCTION_ID } from '../types.js'; const XHS_API = 'https://edith.xiaohongshu.com/api/sns/web/v1'; function buildHeaders(credential: Credential): Record { return { 'Cookie': credential.value, 'User-Agent': 'Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/120.0.0.0 Safari/537.36', 'Referer': 'https://www.xiaohongshu.com', 'Origin': 'https://www.xiaohongshu.com', }; } export class XiaohongshuAdapter implements PlatformAdapter { readonly platformId = 'xiaohongshu'; readonly platformName = '小红书'; async search( keyword: string, _timeRange: string, credential: Credential, _target?: string, ): Promise { try { const params = new URLSearchParams({ keyword, page: '1', page_size: '20', search_id: '', sort: 'general', note_type: '0', }); const response = await fetch(`${XHS_API}/search/notes?${params}`, { headers: buildHeaders(credential), }); if (response.ok) { const data = await response.json() as { code?: number; success?: boolean; data?: { items?: Array<{ id: string; note_card?: { display_title?: string; desc?: string; user?: { nickname?: string }; time?: number; }; }>; }; }; if (data.success && data.data?.items?.length) { return data.data.items.map(item => ({ id: item.id, url: `https://www.xiaohongshu.com/explore/${item.id}`, title: item.note_card?.display_title ?? '', body: item.note_card?.desc ?? '', author: item.note_card?.user?.nickname ?? '', createdAt: item.note_card?.time ? new Date(item.note_card.time * 1000).toISOString() : new Date().toISOString(), platform: this.platformId, })); } } } catch (error) { console.error(`[xiaohongshu] API search failed, falling back to browser: ${(error as Error).message}`); } console.error('[xiaohongshu] API search unavailable, returning browser instruction'); const instruction: BrowserInstruction = { mode: 'browser', action: 'search', steps: [ { action: 'navigate', url: `https://www.xiaohongshu.com/search_result?keyword=${encodeURIComponent(keyword)}&source=web_search_result_notes` }, { action: 'wait', selector: '.note-item' }, { action: 'extract', selector: '.note-item', fields: ['title', 'url', 'author', 'likes'] }, ], cookies: credential.value, }; return [{ id: BROWSER_INSTRUCTION_ID, url: '', title: '', body: JSON.stringify(instruction), author: '', createdAt: new Date().toISOString(), platform: this.platformId, }]; } async reply( postId: string, content: string, credential: Credential, ): Promise { const instruction: BrowserInstruction = { mode: 'browser', action: 'reply', steps: [ { action: 'navigate', url: `https://www.xiaohongshu.com/explore/${postId}` }, { action: 'wait', selector: '#content-textarea' }, { action: 'click', selector: '#content-textarea' }, { action: 'type', text: content }, { action: 'click', selector: '.submit-btn' }, ], cookies: credential.value, }; return { success: true, replyId: `__browser_pending__:${JSON.stringify(instruction)}`, mode: 'browser', }; } async check(credential: Credential): Promise { if (!credential.value || credential.value.length === 0) { return { valid: false, error: 'No cookie provided' }; } return { valid: true, username: '(cookie mode — verify via browser, cookies expire ~12h)', }; } rateLimitInfo(): RateLimitInfo { return { requestsPerMinute: 20, repliesPerDay: 10, minReplyIntervalSeconds: 3, notes: 'No public API. Browser-only. Cookies expire ~12h. Min 3s between requests. Strongly recommend SocialVault for auto cookie refresh.', }; } } const isMainModule = process.argv[1]?.replace(/\\/g, '/').endsWith('xiaohongshu.ts'); if (isMainModule && process.argv[2] === 'test') { const adapter = new XiaohongshuAdapter(); console.log(JSON.stringify({ adapter: 'xiaohongshu', status: 'ok', platformId: adapter.platformId, platformName: adapter.platformName, rateLimit: adapter.rateLimitInfo(), })); } FILE:seeddrop-v3.0.1/scripts/adapters/zhihu.ts // SECURITY MANIFEST: // Environment variables accessed: none // External endpoints called: https://www.zhihu.com/api/v4 // Local files read: none // Local files written: none import type { PlatformAdapter, Credential, Post, ReplyResult, CheckResult, RateLimitInfo, BrowserInstruction, } from '../types.js'; import { BROWSER_INSTRUCTION_ID } from '../types.js'; const ZHIHU_API = 'https://www.zhihu.com/api/v4'; function buildHeaders(credential: Credential): Record { return { 'Cookie': credential.value, 'User-Agent': 'Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/120.0.0.0 Safari/537.36', 'Referer': 'https://www.zhihu.com', }; } export class ZhihuAdapter implements PlatformAdapter { readonly platformId = 'zhihu'; readonly platformName = '知乎'; async search( keyword: string, _timeRange: string, credential: Credential, _target?: string, ): Promise { try { const params = new URLSearchParams({ t: 'general', q: keyword, correction: '1', offset: '0', limit: '20', }); const response = await fetch(`${ZHIHU_API}/search_v3?${params}`, { headers: buildHeaders(credential), }); if (!response.ok) { console.error(`[zhihu] Search failed: ${response.status}`); return []; } const data = await response.json() as { data?: Array<{ type?: string; object?: { id?: number; question?: { id?: number; title?: string }; title?: string; excerpt?: string; content?: string; author?: { name?: string }; created_time?: number; updated_time?: number; }; }>; }; if (!data.data) return []; const posts: Post[] = []; for (const item of data.data) { const obj = item.object; if (!obj) continue; if (item.type === 'search_result' && obj.question) { posts.push({ id: String(obj.question.id ?? obj.id ?? ''), url: `https://www.zhihu.com/question/${obj.question.id}`, title: obj.question.title ?? '', body: (obj.excerpt ?? obj.content ?? '').replace(/<[^>]*>/g, ''), author: obj.author?.name ?? '', createdAt: obj.created_time ? new Date(obj.created_time * 1000).toISOString() : new Date().toISOString(), platform: this.platformId, }); } } const unique = new Map(); for (const p of posts) { if (p.id && !unique.has(p.id)) unique.set(p.id, p); } return Array.from(unique.values()); } catch (error) { console.error(`[zhihu] Search error: ${(error as Error).message}`); return []; } } async reply( postId: string, content: string, credential: Credential, ): Promise { const instruction: BrowserInstruction = { mode: 'browser', action: 'reply', steps: [ { action: 'navigate', url: `https://www.zhihu.com/question/${postId}` }, { action: 'wait', selector: '.AnswerForm' }, { action: 'click', selector: '.AnswerForm .RichContent' }, { action: 'type', text: content }, { action: 'click', selector: 'button[type="submit"]' }, ], cookies: credential.value, }; return { success: true, replyId: `__browser_pending__:${JSON.stringify(instruction)}`, mode: 'browser', }; } async check(credential: Credential): Promise { try { const response = await fetch(`${ZHIHU_API}/me`, { headers: buildHeaders(credential), }); if (!response.ok) return { valid: false, error: `HTTP ${response.status}` }; const data = await response.json() as { id?: string; name?: string; error?: { message?: string } }; if (data.error) { return { valid: false, error: data.error.message ?? 'Unknown error' }; } return { valid: !!data.id, username: data.name }; } catch (error) { return { valid: false, error: (error as Error).message }; } } rateLimitInfo(): RateLimitInfo { return { requestsPerMinute: 10, repliesPerDay: 10, minReplyIntervalSeconds: 300, notes: '知乎: z_c0有效期~30天, 反爬严格, 高频触发验证码. 写入需x-zse-96签名(当前用browser模式)', }; } } const isMainModule = process.argv[1]?.replace(/\\/g, '/').endsWith('zhihu.ts'); if (isMainModule && process.argv[2] === 'test') { const adapter = new ZhihuAdapter(); console.log(JSON.stringify({ adapter: 'zhihu', status: 'ok', platformId: adapter.platformId, platformName: adapter.platformName, rateLimit: adapter.rateLimitInfo(), })); } FILE:seeddrop-v3.0.1/scripts/adapters/_template.ts // SECURITY MANIFEST: // Environment variables accessed: (list here) // External endpoints called: (list here) // Local files read: (list here) // Local files written: (list here) import type { PlatformAdapter, Credential, Post, ReplyResult, CheckResult, RateLimitInfo, } from '../types.js'; /** * Template for creating a new platform adapter. * * Steps: * 1. Copy this file to scripts/adapters/.ts * 2. Rename the class and update platformId / platformName * 3. Implement search(), reply(), check(), rateLimitInfo() * 4. Register in scripts/adapters/base.ts * 5. Create templates/reply-.md * 6. Add rate limits to references/safety-rules.md * 7. Add TOS notes to references/platform-tos-notes.md * 8. Test with: npx tsx scripts/adapters/.ts test */ export class TemplateAdapter implements PlatformAdapter { readonly platformId = 'template'; readonly platformName = 'Template Platform'; async search( keyword: string, timeRange: string, credential: Credential, target?: string, ): Promise { // TODO: Implement platform-specific search // API mode: use fetch() with credential // Browser mode: return BrowserInstruction wrapped in Post[] console.error(`[${this.platformId}] search() not implemented`); return []; } async reply( postId: string, content: string, credential: Credential, ): Promise { // TODO: Implement reply logic console.error(`[${this.platformId}] reply() not implemented`); return { success: false, error: 'Not implemented' }; } async check(credential: Credential): Promise { // TODO: Implement credential validation console.error(`[${this.platformId}] check() not implemented`); return { valid: false, error: 'Not implemented' }; } rateLimitInfo(): RateLimitInfo { return { requestsPerMinute: 60, repliesPerDay: 10, minReplyIntervalSeconds: 300, notes: 'TODO: Add platform-specific rate limit notes', }; } } const isMainModule = process.argv[1]?.replace(/\\/g, '/').endsWith('_template.ts'); if (isMainModule && process.argv[2] === 'test') { const adapter = new TemplateAdapter(); console.log(JSON.stringify({ adapter: 'template', status: 'ok', platformId: adapter.platformId, platformName: adapter.platformName, rateLimit: adapter.rateLimitInfo(), })); } FILE:seeddrop-v3.0.1/references/platform-tos-notes.md # 平台服务条款关键摘要 ## B站 - 搜索接口: `api.bilibili.com/x/web-interface/wbi/search` - 评论接口: `api.bilibili.com/x/v2/reply/add`（需 CSRF=bili_jct） - 搜索频率建议 ≤30 次/小时 - 评论频率建议 ≤50 条/天 - 重复内容会被自动过滤 - 对服务器 IP 无特殊限制 ## 百度贴吧 - BDUSS 有效期 6 个月+，非常稳定 - Cookie-Editor 导出的是 BDUSS_BFESS（不可用），必须从网络请求头获取 BDUSS - 回帖需先获取 tbs token - 回帖频率建议 ≤30 条/天 - 重复内容会被系统删除 - STOKEN 用于写操作验证 ## 知乎 - 搜索接口: `zhihu.com/api/v4/search_v3` - 回答需要 x-zse-96 签名（当前用 browser 模式绕过） - z_c0 有效期约 30 天 - 对服务器 IP 有限制，高频访问触发验证码 - 回答/评论有内容审核 ## 小红书 - 无公开写入 API，回复操作通过浏览器模拟 - 搜索接口: `edith.xiaohongshu.com/api/sns/web/v1/search/notes`（需 X-s/X-t 签名） - Cookie 有效期极短: ~12 小时 - 多层反爬: Header + 行为指纹 + 设备ID + WASM - 强烈建议配合 SocialVault 自动续期 - 评论区不允许外部链接 - 请求间隔不低于 3 秒 FILE:seeddrop-v3.0.1/references/safety-rules.md # SeedDrop 安全规则这些规则在脚本中硬编码执行，不可通过配置绕过。 ## 频率限制 ### 每平台每日回复上限 | 平台 | Approve 模式 | Auto 模式 | 平台限制参考 | |------|-------------|-----------|-------------| | B站 | 30 | 15 | 评论 ≤50/天, 搜索 ≤30/时 | | 贴吧 | 20 | 10 | 回帖 ≤30/天, 重复内容删除 | | 知乎 | 10 | 5 | 反爬严格, 高频触发验证码 | | 小红书 | 10 | 5 | 无 API, ≤1 req/3s | | 其他平台 | 10 | 5 | — | Auto 模式的上限为 Approve 模式的 50%，作为额外安全措施。 ### 回复间隔 - 最小间隔: 5 分钟 - 最大间隔: 15 分钟 - 实际间隔: 在 5-15 分钟范围内随机（不固定） - 小红书额外要求: 任意两次请求间隔不低于 3 秒 ### 同一作者限制 - 同一用户的帖子 24 小时内最多回复 1 次 - 按作者 ID + 平台维度去重 ### 同一帖子限制 - 每个帖子只能回复 1 次 - 通过 `interaction-log.jsonl` 中的 `post_id` 去重 ## 内容安全 - 每条回复必须提供实质性帮助 - 品牌提及不超过回复总内容的 20% - 不使用营销话术、夸大词汇 - 不包含直接联系方式或裸链接 ## 回复多样性 - 不使用固定模板（每次重新生成） - 随机化开头方式、长度、是否提及品牌 ## 禁止回复的场景 - 帖子标记为置顶、官方、管理员帖 - 社区规则明确禁止自我推广 - 帖子作者在黑名单中 - 帖子内容涉及敏感话题 ## Auto 模式额外限制 1. 每日上限降至 Approve 模式的 50% 2. 评分阈值最低不能低于 0.7 3. 风险评分低于 0.5 的帖子不回复 4. 新添加的平台默认 7 天内只能用 Approve 模式 ## 黑名单管理存储在 `memory/blacklist.md`，支持用户/社区/关键词三种维度。 FILE:seeddrop-v3.0.1/references/scoring-criteria.md # SeedDrop 评分标准 ## 评分维度 | 维度 | 权重 | 说明 | |------|------|------| | 相关度 (Relevance) | 35% | 帖子内容与品牌关键词的匹配程度 | | 意图强度 (Intent) | 30% | 帖子作者是否在寻求帮助或建议 | | 时效性 (Freshness) | 20% | 帖子发布时间越新越好 | | 风险评估 (Risk) | 15% | 回复该帖子的潜在风险 | ## 总分计算 ``` final = relevance × 0.35 + intent × 0.30 + freshness × 0.20 + risk × 0.15 ``` ## 阈值 - 默认: 0.6 - Auto 模式最低: 0.7 - 推荐范围: 0.4 - 0.8 ## 相关度评分 | 命中比例 | 分数 | |---------|------| | > 50% 关键词命中 | 1.0 | | 20-50% | 0.7 | | > 0% | 0.5 | | 0% | 0.2 | ## 意图评分 | 类型 | 分数 | 特征词 | |------|------|--------| | 求助/提问 | 0.9 | 求推荐, 怎么, 请问, 有没有推荐, recommend, help | | 讨论/意见 | 0.7 | 讨论, 大家觉得, 有没有人用过, 什么体验 | | 吐槽/抱怨 | 0.5 | 坑, 吐槽, 垃圾, 难用, 避雷 | | 纯分享 | 0.3 | 无特征词 | ## 时效评分 | 帖子年龄 | 分数 | |---------|------| | ≤ 2h | 1.0 | | 2-6h | 0.9 | | 6-12h | 0.8 | | 12-24h | 0.6 | | 24-48h | 0.4 | | > 48h | 0.2 | ## 风险评分 | 类型 | 分数 | |------|------| | 安全帖子 | 0.9 | | 官方/管理相关 | 0.3 | | 争议性话题 | 0.2 | FILE:seeddrop-v3.0.1/memory/blacklist.md # SeedDrop 黑名单 ## 用户黑名单 (不回复以下用户的帖子) ## 社区黑名单 (不在以下社区/subreddit 发布) ## 关键词黑名单 (帖子包含以下关键词时不回复) - [Meta] - [Announcement] - [Mod Post] - [Official] FILE:seeddrop-v3.0.1/memory/brand-profile.md # SeedDrop Brand Profile ## 基本信息 - 业务名称: SeedDrop - 业务类型: 社区互动自动化工具 - 服务区域: 中文互联网（B站、贴吧、知乎、小红书） - 官网: (待配置) - 核心卖点: 自动监控社区讨论，AI生成高质量回复，自然种草，帮助独立开发者和小团队低成本做内容运营 ## 目标客户关键词 - 主关键词: 社区营销, 内容运营工具, 自动回复, 种草工具 - 场景关键词: 独立开发运营, SaaS推广, 小团队营销, 社交媒体运营效率 - 竞品关键词: 社媒管理工具, 内容营销平台 ## 品牌人设 - 语气: 真诚分享、像开发者朋友一样聊天，不硬推销 - 禁用语: 最好的, 第一, 保证, 神器, 爆款, 裂变, 私域 - 常用语: 我们也在用这个方案, 刚好做了个小工具, 分享一下踩过的坑 - 可提及案例: 用 SeedDrop 自动监控「独立开发」「SaaS工具」等话题，每天节省2小时人工筛帖时间 ## 关键词 - 社区营销 - 内容运营工具 - 自动回复 - 种草助手 - 独立开发 - 内容营销 - 运营效率 - SaaS工具推广 - 社交媒体运营 - 社区互动 ## 平台 - bilibili - tieba - zhihu - xiaohongshu ## 监控平台配置 platforms: - id: bilibili enabled: true keywords: [] - id: tieba enabled: true bars: [] keywords: [] - id: zhihu enabled: true keywords: [] - id: xiaohongshu enabled: true keywords: [] ## 运行模式 - mode: approve - scoring_threshold: 0.6 - daily_max_replies: bilibili: 30 tieba: 20 zhihu: 10 xiaohongshu: 10 ## 语言偏好 - primary: zh-CN - secondary: en FILE:seeddrop-v3.0.1/guides/adapter-development.md # 平台适配器开发指南 ## 开发步骤 ### 1. 复制模板 ```bash cp scripts/adapters/_template.ts scripts/adapters/.ts ``` ### 2. 实现 PlatformAdapter 接口 ```typescript import type { PlatformAdapter, Credential, Post, ReplyResult, CheckResult, RateLimitInfo } from '../types.js'; export class MyAdapter implements PlatformAdapter { readonly platformId = 'my-platform'; readonly platformName = 'My Platform'; async search(keyword: string, timeRange: string, credential: Credential, target?: string): Promise { ... } async reply(postId: string, content: string, credential: Credential): Promise { ... } async check(credential: Credential): Promise { ... } rateLimitInfo(): RateLimitInfo { ... } } ``` ### 3. 输出格式 **search 输出**（JSONL，每行一个）： ```json {"id":"...","url":"...","title":"...","body":"...","author":"...","createdAt":"...","platform":""} ``` **reply 输出**： ```json {"success": true, "replyId": "..."} ``` **check 输出**： ```json {"valid": true, "username": "..."} ``` ### 4. 两种实现方式 **有 API 的平台**：使用 `fetch()` 直接调用 API，参考 `reddit.ts`。 **无 API 的平台**：返回 browser 指令，参考 `xiaohongshu.ts`： ```typescript import { BROWSER_INSTRUCTION_ID, type BrowserInstruction } from '../types.js'; const instruction: BrowserInstruction = { mode: 'browser', action: 'search', steps: [ { action: 'navigate', url: `https://example.com/search?q=${keyword}` }, { action: 'wait', selector: '.result-item' }, { action: 'extract', selector: '.result-item', fields: ['title', 'url'] }, ], cookies: credential.value, }; return [{ id: BROWSER_INSTRUCTION_ID, url: '', title: '', body: JSON.stringify(instruction), author: '', createdAt: new Date().toISOString(), platform: this.platformId }]; ``` ### 5. 注册适配器在 `scripts/adapters/base.ts` 中添加： ```typescript 'my-platform': async () => { const { MyAdapter } = await import('./my-platform.js'); return new MyAdapter(); }, ``` ### 6. 配套文件 | 文件 | 用途 | |------|------| | `templates/reply-.md` | 回复风格指南 | | `references/safety-rules.md` | 添加频率限制 | | `references/platform-tos-notes.md` | 添加 ToS 摘要 | ### 7. 测试 ```bash npx tsx scripts/adapters/.ts test ``` ### 8. 启用在 `memory/brand-profile.md` 的 platforms 中添加新平台。 FILE:seeddrop-v3.0.1/guides/brand-profile-setup.md # Brand Profile 配置引导 ## 配置流程执行 `seeddrop setup` 时按以下步骤引导： ### 第 1 步：基本信息 - 业务名称、类型、服务区域 - 官网链接（不主动放入回复） - 核心卖点（1-3 句） ### 第 2 步：关键词 - 主关键词（3-10 个） - 场景关键词（3-10 个） - 竞品关键词（可选） - 中英文都要覆盖 ### 第 3 步：品牌人设 - 语气风格：专业亲切 / 轻松幽默 / 正式严谨 / 热情直接 - 禁用语：如"最好的""保证""第一" - 常用语和可引用案例 ### 第 4 步：平台配置 - 选择启用的平台 - Reddit: 目标 subreddit 列表 - X/Twitter: 话题标签 - 小红书: 搜索关键词 ### 第 5 步：运行模式 - 模式: approve（默认）/ auto - 评分阈值: 默认 0.6 - 每日上限: 各平台独立配置 ### 第 6 步：语言偏好 - 主语言 / 辅助语言 ## 生成文件配置完成写入 `memory/brand-profile.md`，格式参照 `docs/brand-profile.md` 示例。 FILE:seeddrop-v3.0.1/guides/quickstart.md # SeedDrop 快速上手 ## 1. 安装 ```bash clawhub install seeddrop ``` ## 2. 配置品牌资料 ```bash seeddrop setup ``` 按提示输入： - 产品/品牌名称 - 一句话描述 - 监控关键词（多个用逗号分隔） - 使用的平台（bilibili, tieba, zhihu, xiaohongshu） - 回复模式（approve = 人工审核, auto = 自动发送） ## 3. 添加平台凭证 ```bash seeddrop auth add bilibili seeddrop auth add tieba ``` 需要提供对应平台的 Cookie。推荐配合 SocialVault 管理凭证： ```bash clawhub install socialvault socialvault add bilibili ``` ## 4. 运行监控 ```bash seeddrop monitor bilibili seeddrop monitor tieba 编程 seeddrop monitor all ``` ## 5. 查看报告 ```bash seeddrop report seeddrop report weekly ``` ## 平台 Cookie 获取 | 平台 | 必需字段 | 获取方式 | |------|---------|---------| | B站 | SESSDATA, bili_jct | 浏览器 DevTools → Application → Cookies | | 贴吧 | BDUSS, STOKEN | 浏览器 DevTools → Network → 请求头 Cookie | | 知乎 | z_c0, d_c0 | 浏览器 DevTools → Application → Cookies | | 小红书 | a1, web_session | 浏览器 DevTools → Application → Cookies | > **注意**：贴吧的 BDUSS 必须从请求头获取，Cookie-Editor 导出的 BDUSS_BFESS 不可用。 ## 推荐搭配安装 **SocialVault** 后可自动管理 Cookie 续期，特别是小红书（Cookie 仅 12 小时有效）。 FILE:seeddrop-v3.0.1/.qoder/rules/code-quality.md --- trigger: always_on --- # 代码质量与测试标准 ## 命名规范 - 变量名、函数名必须自解释，禁止 temp、data、info、handle 等模糊命名 - 布尔变量用 is/has/should/can 前缀 - 函数用动词开头：calculateTotalPrice 而非 totalPrice - 常量用 UPPER_SNAKE_CASE - 遵循项目已有命名约定（先 grep 确认） ## 函数设计 - 单一职责：一个函数只做一件事 - 函数体不超过 40 行，超过则拆分 - 参数不超过 4 个，超过则用 Options Object - 禁止布尔参数控制分支（用策略模式或独立函数） - 强类型语言中每个函数必须有明确返回值类型 ## 错误处理 - 禁止空 catch 块，每个 catch 必须有日志或重抛 - 外部 I/O 必须有错误处理 - 使用有意义的自定义错误类型 - 异步代码中所有 Promise 必须有错误处理链 - 用户可见错误信息必须友好且不泄漏内部细节 ## 代码结构 - 导入分组：外部库 → 内部模块 → 类型 → 样式 - 相关逻辑物理邻近，不相关逻辑物理分离 - 禁止超过 3 层嵌套（用 early return 或函数抽取） - 魔法数字和魔法字符串必须提取为命名常量 ## 注释哲学 - 好代码本身就是注释，优先靠命名和结构自解释 - 只在需要解释"为什么"时写注释，不解释"是什么" - 公共 API 必须有 JSDoc / docstring - TODO 必须附 issue 链接或负责人 ## 测试标准 ### 验证闭环每次代码变更后必须执行： 1. get_problems → 修到 0 error 2. 运行受影响模块测试 → 全部通过 3. 涉及模块交互时运行集成测试 4. 逐条核对验收标准任何一环不通过则自主修复，禁止标记完成 ### 测试文件生成 - 一次只生成一个测试文件 - 生成后立即 get_problems 检查编译 - 修复编译问题后执行测试 - 当前文件通过后才进入下一个 ### 测试用例设计 - 每个测试用例独立，不依赖执行顺序 - 命名格式：should_[预期行为]_when_[条件] - 必须覆盖：Happy path / Edge cases / Error scenarios / 回归防护 - 每个测试只断言一个行为 - 测试行为和结果，不测实现细节 - Mock 只用于隔离外部依赖 FILE:seeddrop-v3.0.1/.qoder/rules/core-principles.md --- trigger: always_on --- # Vue 2 项目专项规则本规则适用于 Vue 2.x（含 2.7）项目，使用 Options API 风格。 ## 版本识别在执行任何修改前，先确认： - package.json 中 vue 版本为 ^2.x - 构建工具为 Vue CLI (@vue/cli) 或 Webpack - 状态管理为 Vuex（3.x） - UI 库为 Element UI 或其他 Vue 2 兼容库如果发现实际是 Vue 3 项目，立即停止并切换到 Vue 3 规则 ## 组件规范 ### 文件组织 - 单文件组件（SFC）按 template → script → style 顺序排列 - 组件文件名使用 PascalCase：UserProfile.vue - 每个文件只包含一个组件 ### Options API 书写顺序严格按照以下顺序排列 options： 1. name 2. components 3. directives / filters 4. mixins 5. props 6. data 7. computed 8. watch 9. 生命周期钩子（按执行顺序）：beforeCreate → created → beforeMount → mounted → beforeUpdate → updated → beforeDestroy → destroyed 10. methods ### Props 规范 - props 必须有类型定义，禁止纯数组写法 - 正确：props: { title: { type: String, required: true, default: '' } } - 错误：props: ['title'] - 必要时提供 validator 函数 - 复杂对象的 default 必须用工厂函数返回 ### Data 规范 - data 必须是一个返回对象的函数，禁止直接写对象 - data 中不放与模板无关的变量（不需要响应式的数据放在 created 里用 this.xxx = ... 赋值） ### Computed 与 Watch - 能用 computed 解决的不用 watch - watch 中避免复杂逻辑，复杂操作抽到 methods - 需要 deep watch 时显式声明 deep: true 并注释原因 ### 事件规范 - 子组件向父组件通信用 $emit，事件名用 kebab-case - 禁止滥用 $parent / $children 直接访问 - 禁止滥用 EventBus 做跨层级通信（大范围通信走 Vuex） ## Vuex 规范 - State：只存全局共享数据，组件私有状态留在组件内 - Mutations：只做同步状态变更，命名用 SET_XXX / UPDATE_XXX - Actions：处理异步逻辑，命名用动词：fetchUserList / submitOrder - Getters：用于派生状态，等同于 Store 级别的 computed - 模块化：使用 namespaced: true，按业务领域拆分 module ## 路由规范（Vue Router 3.x） - 路由配置使用懒加载：component: () => import('@/views/xxx.vue') - 路由 name 使用 PascalCase，与组件文件名一致 - 路由守卫中的异步操作必须有错误处理 - 禁止在组件内直接操作 window.location，使用 $router ## 样式规范 - 使用 scoped 样式避免全局污染 - 需要穿透子组件时用 ::v-deep 或 /deep/（Vue 2 语法） - 禁止在 scoped style 中使用标签选择器 - BEM 或项目已有的命名约定（先确认再使用） ## 兼容性注意 - Vue 2 不支持 Fragments（模板只能有一个根元素） - Vue 2 不支持 Teleport / Suspense - 数组变更检测限制：使用 Vue.set 或 this.$set 修改数组索引 - 对象新增属性检测限制：使用 Vue.set 或 this.$set 添加新属性 - 如果项目是 Vue 2.7，可使用 Composition API（setup 语法糖除外），但必须先确认项目是否引入了 @vue/composition-api 或已升级到 2.7 ## 禁止事项 - 禁止使用已废弃的 API：$on / $off / $once（Vue 2.x 原生支持但 Vue 3 已移除，迁移风险高） - 禁止在模板中使用复杂的内联表达式（超过一个函数调用的逻辑必须移到 computed 或 methods） - 禁止修改 props（如需修改，用 data 接收或 computed 转换） - 禁止在 data 中引用 props 的默认值后不追踪变化（常见 bug） FILE:seeddrop-v3.0.1/.qoder/rules/documentation-gen.md --- trigger: always_on --- # 文档生成规范通过 @documentation-gen 手动触发。 ## README 模板 - 简介：一句话描述项目用途 - 快速开始：最少步骤跑起来 - 架构概览：核心模块及关系 - 开发指南：环境搭建、常用命令 - API 文档（如适用） - 贡献指南（如适用） ## 变更日志 - 使用 Keep a Changelog 格式 - 按 Added / Changed / Fixed / Removed 分类 - 每条关联 issue 或 PR 编号 ## API 文档 - 包含请求方法、路径、参数、响应示例、错误码 - 提供 curl 示例 FILE:seeddrop-v3.0.1/.qoder/rules/guardrails.md --- trigger: always_on --- # 防翻车护栏机制 ## 认知护栏 ### 不确定性声明对某个技术细节置信度低于 80% 时，必须标注： "⚠️ 我对此不够确定：[内容]。建议验证方式：[方法]" 禁止用自信语气描述不确定的方案。 ### 假设追踪每个假设必须显式记录： "📌 假设：[内容] | 依据：[来源] | 风险：[假设错误的后果]" 验证阶段逐条回顾假设是否成立。 ### 上下文不足时 - 禁止凭"常见做法"猜测项目特定配置 - 必须用工具验证（grep_code 搜模式、read_file 查配置） - 工具也无法确认时向用户明确询问 ## 执行护栏 ### 变更范围锁定 - 每次改动必须与当前任务直接相关 - "顺便改进"的想法记录到 Memory，当前不动 - 任务范围变化时必须更新 Spec 并获得确认 ### 渐进式修改 - 大型修改分步进行，每步之间运行验证 - 禁止一次性重写整个文件，用 search_replace 精确修改 - 每步改完立即 get_problems，确认无回归再继续 ### 并发安全 - 文件编辑：必须串行 - 终端命令：必须串行 - 只读操作（read_file、search_codebase、grep_code）：可以并行 ### 回滚意识 - 高风险操作前确认有回滚路径 - 可能破坏现有功能的修改先确认测试覆盖 - 数据库 migration、配置变更等不可逆操作必须获得用户明确确认 ## 思维链护栏 ### 防止回归倾向 - 面对复杂任务禁止退缩为"建议用户自行处理" - 遇到阻碍先尝试至少 2 种不同解决路径 - 穷尽工具能力后才向用户求助 ### 防止死循环同一修复尝试 3 次仍失败时必须： 1. 停下来重新分析根因 2. 告知用户当前状态和已尝试方案 3. 提出不同方向的解决思路 ### 防止范围蔓延每完成一步回顾：当前在做的事是否仍在 Spec 范围内？发现偏离则立即停止并重新评估。 ## 输出护栏 ### 代码输出 - 默认用 search_replace 修改文件，不直接输出代码块 - 用户明确要求时才在对话中展示代码 - 展示代码用 // ... existing code ... 省略未改部分 ### 任务完成确认任务完成时必须提供结构化报告： - 变更文件清单 - 验收标准逐条达成情况 - 已运行测试及结果 - 潜在风险提示（如有）未实际完成的任务绝不标记为 complete。 FILE:seeddrop-v3.0.1/.qoder/rules/review-and-refactor.md --- trigger: always_on --- # 代码审查与重构 ## 审查清单 ### P0 必须修复 - 逻辑错误：分支遗漏、条件判断错误、循环边界 - 安全漏洞：注入风险、敏感信息泄漏、未授权访问 - 数据一致性：竞态条件、事务缺失、状态不一致 - 资源泄漏：未关闭的连接/文件句柄、内存泄漏 ### P1 强烈建议修复 - 错误处理缺失或不当 - 类型安全问题 - 性能问题（N+1 查询、不必要的重复计算） - 测试覆盖不足 ### P2 改进建议 - 命名可读性 - 代码结构优化 - 注释补充 - 代码重复（DRY） ## 重构原则 - 重构前必须确认测试覆盖 - 一次只做一种类型的重构 - 重构不改变外部行为，每步用测试确认 - 大型重构必须先产出 Spec 获得确认 FILE:seeddrop-v3.0.1/.qoder/rules/security-checklist.md --- trigger: always_on --- # 安全检查清单 ## 输入处理 - 所有用户输入必须验证和消毒（sanitize） - SQL 查询使用参数化查询，禁止字符串拼接 - 文件路径操作防范路径穿越（path traversal） - 正则表达式防范 ReDoS ## 认证授权 - 密码和密钥禁止硬编码 - 敏感配置使用环境变量 - API 端点必须有权限校验 - Session/Token 有合理过期策略 ## 数据安全 - 敏感数据不出现在日志中 - API 响应不泄漏内部实现细节 - 错误信息对用户友好且不暴露技术栈 - PII 数据加密存储 ## 依赖安全 - 不引入有已知漏洞的依赖版本 - 优先使用社区活跃维护的库 FILE:scripts/analytics.ts // SECURITY MANIFEST: // Environment variables accessed: none // External endpoints called: none // Local files read: memory/interaction-log.jsonl, memory/performance-stats.json // Local files written: memory/performance-stats.json, stdout (report) import { readFileSync, writeFileSync, existsSync, mkdirSync } from 'node:fs'; import { join, dirname } from 'node:path'; import { fileURLToPath } from 'node:url'; import type { InteractionLogEntry, PerformanceStats } from './types.js'; const __filename = fileURLToPath(import.meta.url); const __dirname = dirname(__filename); const BASE_DIR = join(__dirname, '..'); const LOG_PATH = join(BASE_DIR, 'memory', 'interaction-log.jsonl'); const STATS_PATH = join(BASE_DIR, 'memory', 'performance-stats.json'); // ─── Data Loading ─────────────────────────────────────────── function loadLog(): InteractionLogEntry[] { if (!existsSync(LOG_PATH)) return []; try { return readFileSync(LOG_PATH, 'utf-8') .split('\n') .filter((l: string) => l.trim()) .map((l: string) => JSON.parse(l) as InteractionLogEntry); } catch (err) { console.error(`[analytics] Failed to load log: ${(err as Error).message}`); return []; } } function loadStats(): PerformanceStats { if (!existsSync(STATS_PATH)) { return { total_replies: 0, by_platform: {}, by_date: {} }; } try { return JSON.parse(readFileSync(STATS_PATH, 'utf-8')) as PerformanceStats; } catch { return { total_replies: 0, by_platform: {}, by_date: {} }; } } function saveStats(stats: PerformanceStats): void { const dir = dirname(STATS_PATH); if (!existsSync(dir)) mkdirSync(dir, { recursive: true }); writeFileSync(STATS_PATH, JSON.stringify(stats, null, 2), 'utf-8'); } // ─── Report Generation ───────────────────────────────────── function filterByDateRange(entries: InteractionLogEntry[], daysBack: number): InteractionLogEntry[] { const cutoff = Date.now() - daysBack * 24 * 60 * 60 * 1000; return entries.filter(e => new Date(e.timestamp).getTime() >= cutoff); } function generateReport(entries: InteractionLogEntry[], label: string): object { const total = entries.length; const successful = entries.filter(e => e.success).length; const failed = total - successful; const byPlatform: Record = {}; const byDate: Record = {}; const byMode: Record = { approve: 0, auto: 0 }; const avgScore = total > 0 ? entries.reduce((sum, e) => sum + e.score, 0) / total : 0; for (const entry of entries) { const plat = entry.platform; if (!byPlatform[plat]) byPlatform[plat] = { total: 0, success: 0 }; byPlatform[plat].total++; if (entry.success) byPlatform[plat].success++; const dateKey = entry.timestamp.substring(0, 10); byDate[dateKey] = (byDate[dateKey] ?? 0) + 1; byMode[entry.mode] = (byMode[entry.mode] ?? 0) + 1; } return { report: label, period: { from: entries.length > 0 ? entries[0].timestamp : null, to: entries.length > 0 ? entries[entries.length - 1].timestamp : null, }, summary: { total, successful, failed, successRate: total > 0 ? Math.round((successful / total) * 100) : 0, averageScore: Math.round(avgScore * 1000) / 1000, }, byPlatform, byDate, byMode, }; } function generateTuningAdvice(entries: InteractionLogEntry[]): object { const suggestions: string[] = []; if (entries.length === 0) { return { suggestions: ['No data yet. Run some monitoring cycles first.'] }; } const avgScore = entries.reduce((s, e) => s + e.score, 0) / entries.length; if (avgScore < 0.6) { suggestions.push('Average score is low. Consider refining keywords in brand-profile.md to target more relevant discussions.'); } if (avgScore > 0.85) { suggestions.push('Average score is very high. You might lower the threshold slightly to capture more opportunities.'); } const successRate = entries.filter(e => e.success).length / entries.length; if (successRate < 0.7) { suggestions.push('Success rate is below 70%. Check credential validity and platform rate limits.'); } const platformCounts: Record = {}; for (const e of entries) { platformCounts[e.platform] = (platformCounts[e.platform] ?? 0) + 1; } const platforms = Object.keys(platformCounts); if (platforms.length === 1) { suggestions.push(`All activity is on ${platforms[0]}. Consider expanding to other platforms for broader reach.`); } const hourBuckets = new Array(24).fill(0); for (const e of entries) { const hour = new Date(e.timestamp).getHours(); hourBuckets[hour]++; } const peakHour = hourBuckets.indexOf(Math.max(...hourBuckets)); suggestions.push(`Peak activity hour: ${peakHour}:00. Consider scheduling monitoring around this time.`); if (suggestions.length === 0) { suggestions.push('Everything looks good! Keep the current configuration.'); } return { tuning: 'advice', averageScore: Math.round(avgScore * 1000) / 1000, successRate: Math.round(successRate * 100), platformDistribution: platformCounts, peakHour, suggestions, }; } // ─── Stats Update ─────────────────────────────────────────── function updateStats(entries: InteractionLogEntry[]): void { const stats = loadStats(); stats.total_replies = entries.filter(e => e.success).length; stats.by_platform = {}; stats.by_date = {}; for (const entry of entries) { if (!entry.success) continue; stats.by_platform[entry.platform] = (stats.by_platform[entry.platform] ?? 0) + 1; const dateKey = entry.timestamp.substring(0, 10); stats.by_date[dateKey] = (stats.by_date[dateKey] ?? 0) + 1; } saveStats(stats); console.error('[analytics] performance-stats.json updated'); } // ─── CLI Entry Point ──────────────────────────────────────── function main(): void { const args = process.argv.slice(2); const command = args[0] ?? 'daily'; if (command === 'test') { const log = loadLog(); const stats = loadStats(); console.log(JSON.stringify({ script: 'analytics', status: 'ok', logEntries: log.length, currentStats: stats, })); return; } const log = loadLog(); switch (command) { case 'daily': { const today = filterByDateRange(log, 1); const report = generateReport(today, 'daily'); updateStats(log); console.log(JSON.stringify(report, null, 2)); break; } case 'weekly': { const week = filterByDateRange(log, 7); const report = generateReport(week, 'weekly'); updateStats(log); console.log(JSON.stringify(report, null, 2)); break; } case 'tune': { const recent = filterByDateRange(log, 14); const advice = generateTuningAdvice(recent); console.log(JSON.stringify(advice, null, 2)); break; } default: console.error('Usage: analytics.ts '); process.exit(1); } } main(); FILE:scripts/auth-bridge.ts // SECURITY MANIFEST: // Environment variables accessed: HOME, USERPROFILE // External endpoints called: none (SocialVault calls delegated to Agent) // Local files read: SocialVault SKILL.md (existence check only) // Local files written: none // Security: SocialVault is REQUIRED - no plaintext credential fallback import { existsSync } from 'node:fs'; import { homedir } from 'node:os'; import { join, dirname } from 'node:path'; import { fileURLToPath } from 'node:url'; import type { Credential, AuthMode, CheckResult } from './types.js'; const __filename = fileURLToPath(import.meta.url); const __dirname = dirname(__filename); const BASE_DIR = join(__dirname, '..'); // ─── SocialVault Detection ────────────────────────────────── const SOCIALVAULT_SEARCH_PATHS = [ join(homedir(), '.openclaw', 'skills', 'socialvault', 'SKILL.md'), join(homedir(), '.openclaw', 'skills', 'social-vault', 'SKILL.md'), join(homedir(), '.openclaw', 'workspace', 'skills', 'socialvault', 'SKILL.md'), join(homedir(), '.openclaw', 'workspace', 'skills', 'social-vault', 'SKILL.md'), join(BASE_DIR, '..', 'socialvault', 'SKILL.md'), join(BASE_DIR, '..', 'social-vault', 'SKILL.md'), join(BASE_DIR, '..', 'SocialVault', 'socialvault', 'SKILL.md'), ]; function detectSocialVault(): string | null { for (const p of SOCIALVAULT_SEARCH_PATHS) { if (existsSync(p)) { console.error(`[auth-bridge] SocialVault detected at: ${dirname(p)}`); return p; } } return null; } export function getAuthMode(): AuthMode { return detectSocialVault() ? 'socialvault' : 'none'; } // ─── SocialVault Mode ─────────────────────────────────────── function getSocialVaultInstruction(command: string, platform: string, profile: string): string { switch (command) { case 'use': return `socialvault use ${platform}-${profile}`; case 'token': return `socialvault token ${platform}-${profile}`; case 'release': return `socialvault release ${platform}-${profile}`; case 'check': return `socialvault check ${platform}-${profile}`; default: return `socialvault status`; } } // ─── Public API ───────────────────────────────────────────── export function getCredential(platform: string, profile: string = 'default'): Credential | null { const mode = getAuthMode(); if (mode === 'socialvault') { const instruction = getSocialVaultInstruction('token', platform, profile); console.error(`[auth-bridge] SocialVault mode — Agent should run: ${instruction}`); return { authType: 'oauth', value: `__socialvault_pending__:${instruction}`, profile, source: 'socialvault', }; } console.error(`[auth-bridge] SocialVault is required but not detected. Please install: clawhub install socialvault`); return null; } export function checkCredential(platform: string, profile: string = 'default'): CheckResult { const cred = getCredential(platform, profile); if (!cred) { return { valid: false, error: `No credential found for ${platform}/${profile}` }; } if (cred.source === 'socialvault') { return { valid: true, username: `(via SocialVault, run: socialvault check ${platform}-${profile})` }; } return { valid: cred.value.length > 0, error: cred.value.length === 0 ? 'Credential value is empty' : undefined, }; } // ─── CLI Entry Point ──────────────────────────────────────── const IS_MAIN = process.argv[1]?.replace(/\\/g, '/').endsWith('auth-bridge.ts'); function main(): void { if (!IS_MAIN) return; const args = process.argv.slice(2); const command = args[0]; switch (command) { case 'mode': console.log(JSON.stringify({ mode: getAuthMode() })); break; case 'get': { const platform = args[1]; const profile = args[2] ?? 'default'; if (!platform) { console.error('Usage: auth-bridge.ts get [profile]'); process.exit(1); } const cred = getCredential(platform, profile); if (!cred) { console.error(`[auth-bridge] Failed to get credential for ${platform}/${profile}`); process.exit(1); } console.log(JSON.stringify(cred)); break; } case 'check': { const platform = args[1]; const profile = args[2] ?? 'default'; if (!platform) { console.error('Usage: auth-bridge.ts check [profile]'); process.exit(1); } const result = checkCredential(platform, profile); console.log(JSON.stringify(result)); break; } case 'test': console.log(JSON.stringify({ script: 'auth-bridge', status: 'ok', mode: getAuthMode(), socialVaultDetected: detectSocialVault() !== null, })); break; default: console.error('Usage: auth-bridge.ts [args]'); console.error(' mode — Show auth mode (local/socialvault)'); console.error(' get [profile] — Get credential'); console.error(' check [profile] — Check credential validity'); console.error(' test — Self-test'); process.exit(1); } } main(); FILE:scripts/monitor.ts // SECURITY MANIFEST: // Environment variables accessed: none // External endpoints called: none (delegated to adapters) // Local files read: memory/brand-profile.md, memory/interaction-log.jsonl, memory/blacklist.md // Local files written: stdout (JSONL) import { readFileSync, existsSync } from 'node:fs'; import { join, dirname } from 'node:path'; import { fileURLToPath } from 'node:url'; import { getAdapter, listPlatforms } from './adapters/base.js'; import { getCredential } from './auth-bridge.js'; import type { Post, InteractionLogEntry } from './types.js'; const __filename = fileURLToPath(import.meta.url); const __dirname = dirname(__filename); const BASE_DIR = join(__dirname, '..'); // ─── Brand Profile ────────────────────────────────────────── interface MonitorConfig { keywords: string[]; platforms: string[]; timeRange: string; } function loadConfig(): MonitorConfig { const profilePath = join(BASE_DIR, 'memory', 'brand-profile.md'); const defaults: MonitorConfig = { keywords: [], platforms: [], timeRange: 'day', }; if (!existsSync(profilePath)) { console.error('[monitor] brand-profile.md not found'); return defaults; } try { const raw = readFileSync(profilePath, 'utf-8'); const kwMatch = raw.match(/##\s*关键词.*?\n([\s\S]*?)(?=\n##|\n$|$)/i) ?? raw.match(/##\s*Keywords.*?\n([\s\S]*?)(?=\n##|\n$|$)/i); if (kwMatch) { defaults.keywords = kwMatch[1] .split('\n') .map((l: string) => l.replace(/^[-*]\s*/, '').trim()) .filter((l: string) => l.length > 0); } const platMatch = raw.match(/##\s*平台.*?\n([\s\S]*?)(?=\n##|\n$|$)/i) ?? raw.match(/##\s*Platforms.*?\n([\s\S]*?)(?=\n##|\n$|$)/i); if (platMatch) { defaults.platforms = platMatch[1] .split('\n') .map((l: string) => l.replace(/^[-*]\s*/, '').trim()) .filter((l: string) => l.length > 0); } return defaults; } catch (err) { console.error(`[monitor] Failed to load config: ${(err as Error).message}`); return defaults; } } // ─── Deduplication ────────────────────────────────────────── function loadRepliedPostIds(): Set { const logPath = join(BASE_DIR, 'memory', 'interaction-log.jsonl'); if (!existsSync(logPath)) return new Set(); try { const raw = readFileSync(logPath, 'utf-8'); const ids = new Set(); for (const line of raw.split('\n')) { if (!line.trim()) continue; try { const entry = JSON.parse(line) as InteractionLogEntry; ids.add(entry.postId); } catch { /* skip malformed lines */ } } return ids; } catch { return new Set(); } } // ─── Blacklist ────────────────────────────────────────────── interface Blacklist { users: string[]; communities: string[]; keywords: string[]; } function loadBlacklist(): Blacklist { const path = join(BASE_DIR, 'memory', 'blacklist.md'); const defaults: Blacklist = { users: [], communities: [], keywords: [] }; if (!existsSync(path)) return defaults; try { const raw = readFileSync(path, 'utf-8'); let currentSection = ''; for (const line of raw.split('\n')) { const sectionMatch = line.match(/^##\s*(.+)/); if (sectionMatch) { currentSection = sectionMatch[1].toLowerCase(); continue; } const item = line.replace(/^[-*]\s*/, '').trim(); if (!item || item.startsWith('#')) continue; if (currentSection.includes('用户') || currentSection.includes('user')) { defaults.users.push(item.toLowerCase()); } else if (currentSection.includes('社区') || currentSection.includes('communit') || currentSection.includes('subreddit')) { defaults.communities.push(item.toLowerCase()); } else if (currentSection.includes('关键词') || currentSection.includes('keyword')) { defaults.keywords.push(item.toLowerCase()); } } return defaults; } catch { return defaults; } } function isBlacklisted(post: Post, blacklist: Blacklist): boolean { if (blacklist.users.includes(post.author.toLowerCase())) return true; if (post.community && blacklist.communities.includes(post.community.toLowerCase())) return true; const text = `${post.title} ${post.body}`.toLowerCase(); for (const kw of blacklist.keywords) { if (text.includes(kw)) return true; } return false; } // ─── Main Monitor Logic ───────────────────────────────────── async function monitorPlatform( platformId: string, keywords: string[], timeRange: string, target?: string, ): Promise { const adapter = await getAdapter(platformId); const cred = getCredential(platformId); if (!cred) { console.error(`[monitor] No credentials for ${platformId}, skipping`); return []; } if (cred.value.startsWith('__socialvault_pending__:')) { console.error(`[monitor] SocialVault credential pending for ${platformId}`); console.error(`[monitor] Agent should run: ${cred.value.replace('__socialvault_pending__:', '')}`); console.error(`[monitor] Then re-run monitor with valid credentials`); return []; } const allPosts: Post[] = []; for (const keyword of keywords) { console.error(`[monitor] Searching ${platformId} for "${keyword}" (range: ${timeRange})${target ? ` in ${target}` : ''}`); const posts = await adapter.search(keyword, timeRange, cred, target); if (posts.length === 0 && adapter.browserSearch) { const instruction = adapter.browserSearch(keyword, target); console.error(`[monitor] API search returned 0 results — browser fallback available.`); console.error(`[monitor] BROWSER_FALLBACK: ${JSON.stringify(instruction)}`); } allPosts.push(...posts); } const uniquePosts = new Map(); for (const post of allPosts) { if (!uniquePosts.has(post.id)) { uniquePosts.set(post.id, post); } } return Array.from(uniquePosts.values()); } // ─── CLI Entry Point ──────────────────────────────────────── async function main(): Promise { const args = process.argv.slice(2); if (args[0] === 'test') { const config = loadConfig(); const blacklist = loadBlacklist(); const repliedIds = loadRepliedPostIds(); console.log(JSON.stringify({ script: 'monitor', status: 'ok', config, blacklistCounts: { users: blacklist.users.length, communities: blacklist.communities.length, keywords: blacklist.keywords.length, }, repliedPostCount: repliedIds.size, availablePlatforms: listPlatforms(), })); return; } const platformArg = args[0]; const targetArg = args[1]; // e.g. specific community (吧名, 分区等) if (!platformArg) { console.error('Usage: monitor.ts [target]'); console.error(`Available platforms: ${listPlatforms().join(', ')}`); process.exit(1); } const config = loadConfig(); if (config.keywords.length === 0) { console.error('[monitor] No keywords configured in brand-profile.md'); console.error('[monitor] Run "seeddrop setup" to configure keywords'); process.exit(1); } const blacklist = loadBlacklist(); const repliedIds = loadRepliedPostIds(); const platforms = platformArg === 'all' ? (config.platforms.length > 0 ? config.platforms : listPlatforms()) : [platformArg]; let totalOutput = 0; for (const plat of platforms) { try { const posts = await monitorPlatform(plat, config.keywords, config.timeRange, targetArg); console.error(`[monitor] ${plat}: found ${posts.length} raw posts`); for (const post of posts) { if (repliedIds.has(post.id)) { console.error(`[monitor] SKIP ${post.id}: already replied`); continue; } if (isBlacklisted(post, blacklist)) { console.error(`[monitor] SKIP ${post.id}: blacklisted`); continue; } console.log(JSON.stringify(post)); totalOutput++; } } catch (err) { console.error(`[monitor] Error monitoring ${plat}: ${(err as Error).message}`); } } console.error(`[monitor] Total output: ${totalOutput} posts`); } main().catch(err => { console.error(`[monitor] Fatal error: ${(err as Error).message}`); process.exit(1); }); FILE:scripts/responder.ts // SECURITY MANIFEST: // Environment variables accessed: none // External endpoints called: none (reply sending delegated to adapters via Agent) // Local files read: stdin (JSONL), memory/interaction-log.jsonl, memory/brand-profile.md // Local files written: stdout (drafts/results), memory/interaction-log.jsonl import { readFileSync, appendFileSync, existsSync, mkdirSync } from 'node:fs'; import { join, dirname } from 'node:path'; import { fileURLToPath } from 'node:url'; import { createInterface } from 'node:readline'; import type { ScoredPost, InteractionLogEntry, ReplyDraft, ReplyMode, } from './types.js'; import { PLATFORM_DAILY_LIMITS } from './types.js'; const __filename = fileURLToPath(import.meta.url); const __dirname = dirname(__filename); const BASE_DIR = join(__dirname, '..'); const LOG_PATH = join(BASE_DIR, 'memory', 'interaction-log.jsonl'); // ─── Interaction Log ──────────────────────────────────────── function loadInteractionLog(): InteractionLogEntry[] { if (!existsSync(LOG_PATH)) return []; try { const raw = readFileSync(LOG_PATH, 'utf-8'); return raw .split('\n') .filter((l: string) => l.trim()) .map((l: string) => JSON.parse(l) as InteractionLogEntry); } catch (err) { console.error(`[responder] Failed to load interaction log: ${(err as Error).message}`); return []; } } export function appendToLog(entry: InteractionLogEntry): void { const dir = dirname(LOG_PATH); if (!existsSync(dir)) mkdirSync(dir, { recursive: true }); appendFileSync(LOG_PATH, JSON.stringify(entry) + '\n', 'utf-8'); } // ─── Deduplication & Safety Checks ────────────────────────── function isAlreadyReplied(postId: string, log: InteractionLogEntry[]): boolean { return log.some(entry => entry.postId === postId); } function isAuthorCoolingDown(author: string, platform: string, log: InteractionLogEntry[]): boolean { const cutoff = Date.now() - 24 * 60 * 60 * 1000; return log.some( entry => entry.author === author && entry.platform === platform && new Date(entry.timestamp).getTime() > cutoff, ); } function getTodayReplyCount(platform: string, log: InteractionLogEntry[]): number { const todayStart = new Date(); todayStart.setHours(0, 0, 0, 0); const cutoff = todayStart.getTime(); return log.filter( entry => entry.platform === platform && entry.success && new Date(entry.timestamp).getTime() >= cutoff, ).length; } function getDailyLimit(platform: string, mode: ReplyMode): number { const limits = PLATFORM_DAILY_LIMITS[platform] ?? PLATFORM_DAILY_LIMITS['_default']!; return mode === 'auto' ? limits.auto : limits.approve; } // ─── Reply Draft Generation ───────────────────────────────── function generateDraft(post: ScoredPost): ReplyDraft { return { postId: post.id, postUrl: post.url, postTitle: post.title, platform: post.platform, content: `__DRAFT_PLACEHOLDER__`, score: post.finalScore, }; } // ─── CLI Entry Point ──────────────────────────────────────── async function main(): Promise { const args = process.argv.slice(2); if (args[0] === 'test') { const log = loadInteractionLog(); console.log(JSON.stringify({ script: 'responder', status: 'ok', logEntries: log.length, logPath: LOG_PATH, })); return; } // SECURITY: Only approve mode is supported - no auto-reply to prevent abuse const mode: ReplyMode = 'approve'; if (args[0] === 'auto') { console.error('[responder] SECURITY: Auto mode is disabled. All replies require manual approval.'); process.exit(1); } console.error(`[responder] Mode: ${mode} (manual approval required)`); const log = loadInteractionLog(); const rl = createInterface({ input: process.stdin, terminal: false }); let inputCount = 0; let draftCount = 0; let skipDup = 0; let skipAuthor = 0; let skipLimit = 0; for await (const line of rl) { if (!line.trim()) continue; try { const post = JSON.parse(line) as ScoredPost; inputCount++; if (isAlreadyReplied(post.id, log)) { console.error(`[responder] SKIP ${post.id}: already replied`); skipDup++; continue; } if (isAuthorCoolingDown(post.author, post.platform, log)) { console.error(`[responder] SKIP ${post.id}: author "${post.author}" cooldown (24h)`); skipAuthor++; continue; } const dailyLimit = getDailyLimit(post.platform, mode); const todayCount = getTodayReplyCount(post.platform, log); if (todayCount >= dailyLimit) { console.error(`[responder] SKIP ${post.id}: daily limit reached (${todayCount}/${dailyLimit})`); skipLimit++; continue; } const draft = generateDraft(post); if (mode === 'approve') { console.log(JSON.stringify({ action: 'review_draft', draft, post: { id: post.id, url: post.url, title: post.title, body: post.body.substring(0, 300), author: post.author, platform: post.platform, community: post.community, scores: post.scores, finalScore: post.finalScore, }, instructions: `Generate a value-first reply for this ${post.platform} post. Follow the template at templates/reply-${post.platform}.md. Brand mention ≤20%. Then present the draft to the user for approval.`, })); } else { console.log(JSON.stringify({ action: 'auto_reply', draft, post: { id: post.id, url: post.url, title: post.title, body: post.body.substring(0, 300), author: post.author, platform: post.platform, community: post.community, scores: post.scores, finalScore: post.finalScore, }, instructions: `Generate and send a value-first reply for this ${post.platform} post. Follow the template at templates/reply-${post.platform}.md. Brand mention ≤20%. Log the result to interaction-log.jsonl.`, })); } draftCount++; } catch { console.error(`[responder] Failed to parse line: ${line.substring(0, 80)}`); } } console.error(`[responder] Done: ${draftCount} drafts from ${inputCount} posts (skipped: ${skipDup} dup, ${skipAuthor} author, ${skipLimit} limit)`); } main().catch(err => { console.error(`[responder] Fatal error: ${(err as Error).message}`); process.exit(1); }); FILE:scripts/scorer.ts // SECURITY MANIFEST: // Environment variables accessed: none // External endpoints called: none // Local files read: stdin (JSONL), memory/brand-profile.md // Local files written: stdout (JSONL) import { readFileSync, existsSync } from 'node:fs'; import { join, dirname } from 'node:path'; import { fileURLToPath } from 'node:url'; import { createInterface } from 'node:readline'; import type { Post, ScoredPost, ScoreBreakdown } from './types.js'; import { SCORE_WEIGHTS, DEFAULT_THRESHOLD, AUTO_MODE_MIN_THRESHOLD, AUTO_MODE_MIN_RISK, } from './types.js'; const __filename = fileURLToPath(import.meta.url); const __dirname = dirname(__filename); const BASE_DIR = join(__dirname, '..'); // ─── Brand Profile Loading ────────────────────────────────── interface ParsedBrandProfile { keywords: string[]; mode: 'approve' | 'auto'; threshold: number; } function loadBrandProfile(): ParsedBrandProfile { const profilePath = join(BASE_DIR, 'memory', 'brand-profile.md'); const defaults: ParsedBrandProfile = { keywords: [], mode: 'approve', threshold: DEFAULT_THRESHOLD, }; if (!existsSync(profilePath)) { console.error('[scorer] brand-profile.md not found, using defaults'); return defaults; } try { const raw = readFileSync(profilePath, 'utf-8'); const keywordsMatch = raw.match(/##\s*关键词.*?\n([\s\S]*?)(?=\n##|\n$|$)/i) ?? raw.match(/##\s*Keywords.*?\n([\s\S]*?)(?=\n##|\n$|$)/i); if (keywordsMatch) { defaults.keywords = keywordsMatch[1] .split('\n') .map((l: string) => l.replace(/^[-*]\s*/, '').trim()) .filter((l: string) => l.length > 0); } const modeMatch = raw.match(/模式[：:]\s*(approve|auto)/i) ?? raw.match(/mode[：:]\s*(approve|auto)/i); if (modeMatch) { defaults.mode = modeMatch[1].toLowerCase() as 'approve' | 'auto'; } const thresholdMatch = raw.match(/阈值[：:]\s*([\d.]+)/i) ?? raw.match(/threshold[：:]\s*([\d.]+)/i); if (thresholdMatch) { defaults.threshold = parseFloat(thresholdMatch[1]); } return defaults; } catch (err) { console.error(`[scorer] Failed to load brand profile: ${(err as Error).message}`); return defaults; } } // ─── Scoring Functions ────────────────────────────────────── const INTENT_KEYWORDS = { high: ['求推荐', '怎么', '有什么好的', '推荐一下', '求助', '请问', '有没有推荐', '怎么选', '怎么办', 'how to', 'recommend', 'help', 'looking for'], medium: ['讨论', '大家觉得', '有没有人用过', '有经验', '分享一下', '什么体验', '值不值', 'what do you think', 'experience'], low: ['坑', '吐槽', '垃圾', '难用', '太差了', '后悔', '避雷', 'frustrated', 'terrible'], }; function scoreRelevance(post: Post, keywords: string[]): number { if (keywords.length === 0) return 0.5; const text = `${post.title} ${post.body}`.toLowerCase(); let hits = 0; for (const kw of keywords) { if (text.includes(kw.toLowerCase())) hits++; } const ratio = hits / keywords.length; if (ratio > 0.5) return 1.0; if (ratio > 0.2) return 0.7; if (ratio > 0) return 0.5; return 0.2; } function scoreIntent(post: Post): number { const text = `${post.title} ${post.body}`.toLowerCase(); for (const kw of INTENT_KEYWORDS.high) { if (text.includes(kw)) return 0.9; } for (const kw of INTENT_KEYWORDS.medium) { if (text.includes(kw)) return 0.7; } for (const kw of INTENT_KEYWORDS.low) { if (text.includes(kw)) return 0.5; } return 0.3; } function scoreFreshness(post: Post): number { const ageMs = Date.now() - new Date(post.createdAt).getTime(); const ageHours = ageMs / (1000 * 60 * 60); if (ageHours <= 2) return 1.0; if (ageHours <= 6) return 0.9; if (ageHours <= 12) return 0.8; if (ageHours <= 24) return 0.6; if (ageHours <= 48) return 0.4; return 0.2; } function scoreRisk(post: Post): number { const text = `${post.title} ${post.body}`.toLowerCase(); const officialMarkers = ['[meta]', '[announcement]', '[mod]', '[official]', '置顶', 'pinned']; for (const marker of officialMarkers) { if (text.includes(marker)) return 0.3; } const controversialMarkers = ['politic', 'religion', 'controversial', '政治', '宗教', '敏感']; for (const marker of controversialMarkers) { if (text.includes(marker)) return 0.2; } return 0.9; } export function scorePost(post: Post, keywords: string[]): ScoredPost { const scores: ScoreBreakdown = { relevance: scoreRelevance(post, keywords), intent: scoreIntent(post), freshness: scoreFreshness(post), risk: scoreRisk(post), }; const finalScore = scores.relevance * SCORE_WEIGHTS.relevance + scores.intent * SCORE_WEIGHTS.intent + scores.freshness * SCORE_WEIGHTS.freshness + scores.risk * SCORE_WEIGHTS.risk; return { ...post, scores, finalScore: Math.round(finalScore * 1000) / 1000 }; } // ─── CLI Entry Point ──────────────────────────────────────── async function main(): Promise { const args = process.argv.slice(2); if (args[0] === 'test') { const testPost: Post = { id: 'test1', url: 'https://reddit.com/r/test/test1', title: 'How to build a landing page?', body: 'Looking for recommendations on tools to build a quick landing page.', author: 'testuser', createdAt: new Date().toISOString(), platform: 'reddit', }; const result = scorePost(testPost, ['landing page', 'build', 'tool']); console.log(JSON.stringify({ script: 'scorer', status: 'ok', testResult: result, })); return; } const thresholdArg = parseFloat(args[0] || ''); const profile = loadBrandProfile(); const threshold = !isNaN(thresholdArg) ? thresholdArg : profile.threshold; const effectiveThreshold = profile.mode === 'auto' ? Math.max(threshold, AUTO_MODE_MIN_THRESHOLD) : threshold; console.error(`[scorer] Mode: ${profile.mode}, Threshold: ${effectiveThreshold}, Keywords: ${profile.keywords.length}`); const rl = createInterface({ input: process.stdin, terminal: false }); let inputCount = 0; let passedCount = 0; for await (const line of rl) { if (!line.trim()) continue; try { const post = JSON.parse(line) as Post; inputCount++; const scored = scorePost(post, profile.keywords); if (scored.finalScore < effectiveThreshold) { console.error(`[scorer] SKIP ${scored.id} (score=${scored.finalScore} < ${effectiveThreshold})`); continue; } if (profile.mode === 'auto' && scored.scores.risk < AUTO_MODE_MIN_RISK) { console.error(`[scorer] SKIP ${scored.id} (risk=${scored.scores.risk} < ${AUTO_MODE_MIN_RISK}, auto mode)`); continue; } passedCount++; console.log(JSON.stringify(scored)); } catch { console.error(`[scorer] Failed to parse line: ${line.substring(0, 80)}`); } } console.error(`[scorer] Done: ${passedCount}/${inputCount} posts passed threshold`); } main().catch(err => { console.error(`[scorer] Fatal error: ${(err as Error).message}`); process.exit(1); }); FILE:scripts/types.ts // SECURITY MANIFEST: // Environment variables accessed: none // External endpoints called: none // Local files read: none // Local files written: none // ─── Credential ───────────────────────────────────────────── export interface Credential { authType: 'api_token' | 'cookie' | 'oauth'; value: string; profile?: string; source: 'socialvault' | 'local'; } // ─── Post & Scoring ───────────────────────────────────────── export interface Post { id: string; url: string; title: string; body: string; author: string; createdAt: string; // ISO 8601 platform: string; community?: string; // 吧名、知乎话题、B站分区等 metadata?: Record; } export interface ScoreBreakdown { relevance: number; intent: number; freshness: number; risk: number; } export interface ScoredPost extends Post { scores: ScoreBreakdown; finalScore: number; } // ─── Reply ────────────────────────────────────────────────── export interface ReplyResult { success: boolean; replyId?: string; error?: string; mode?: 'api' | 'browser'; } export interface ReplyDraft { postId: string; postUrl: string; postTitle: string; platform: string; content: string; score: number; } // ─── Auth / Check ─────────────────────────────────────────── export interface CheckResult { valid: boolean; username?: string; error?: string; } export type AuthMode = 'socialvault' | 'none'; // ─── Rate Limiting ────────────────────────────────────────── export interface RateLimitInfo { requestsPerMinute: number; repliesPerDay: number; minReplyIntervalSeconds: number; notes: string; } export interface DailyLimits { approve: number; auto: number; } export const PLATFORM_DAILY_LIMITS: Record = { 'bilibili': { approve: 30, auto: 15 }, 'tieba': { approve: 20, auto: 10 }, 'zhihu': { approve: 10, auto: 5 }, 'xiaohongshu': { approve: 10, auto: 5 }, '_default': { approve: 10, auto: 5 }, }; // ─── Platform Adapter ─────────────────────────────────────── export interface PlatformAdapter { readonly platformId: string; readonly platformName: string; search( keyword: string, timeRange: string, credential: Credential, target?: string, ): Promise; reply( postId: string, content: string, credential: Credential, ): Promise; check(credential: Credential): Promise; rateLimitInfo(): RateLimitInfo; /** * Returns BrowserInstruction for Agent to execute search via browser tool. * Used as fallback when API search is blocked (403). */ browserSearch?( keyword: string, target?: string, ): BrowserInstruction; } // ─── Interaction Log ──────────────────────────────────────── export interface InteractionLogEntry { timestamp: string; // ISO 8601 platform: string; postId: string; postUrl: string; postTitle: string; author: string; replyContent: string; replyId?: string; score: number; mode: 'approve' | 'auto'; success: boolean; } // ─── Brand Profile ────────────────────────────────────────── export interface BrandProfile { businessName: string; description: string; keywords: string[]; platforms: string[]; mode: 'approve' | 'auto'; threshold: number; language: string; } // ─── Scoring Config ───────────────────────────────────────── export const SCORE_WEIGHTS = { relevance: 0.35, intent: 0.30, freshness: 0.20, risk: 0.15, } as const; export const DEFAULT_THRESHOLD = 0.6; export const AUTO_MODE_MIN_THRESHOLD = 0.7; export const AUTO_MODE_MIN_RISK = 0.5; // ─── Browser Instruction (for adapter browser mode) ───────── export const BROWSER_INSTRUCTION_ID = '__browser_instruction__'; export interface BrowserStep { action: 'navigate' | 'wait' | 'extract' | 'click' | 'type'; url?: string; selector?: string; fields?: string[]; text?: string; } export interface BrowserInstruction { mode: 'browser'; action: 'search' | 'reply' | 'check'; steps: BrowserStep[]; cookies?: string; } // ─── Anti-Detection Helpers ────────────────────────────────── const CHROME_VERSION = '131'; export function buildBrowserHeaders(cookie: string, extra?: Record): Record { return { 'Cookie': cookie, 'User-Agent': `Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/${CHROME_VERSION}.0.0.0 Safari/537.36`, 'Accept': 'text/html,application/xhtml+xml,application/xml;q=0.9,image/avif,image/webp,image/apng,*/*;q=0.8', 'Accept-Language': 'zh-CN,zh;q=0.9,en;q=0.8', 'Accept-Encoding': 'gzip, deflate, br', 'Cache-Control': 'max-age=0', 'Sec-Fetch-Dest': 'document', 'Sec-Fetch-Mode': 'navigate', 'Sec-Fetch-Site': 'none', 'Sec-Fetch-User': '?1', 'Sec-Ch-Ua': `"Chromium";v="${CHROME_VERSION}", "Not_A Brand";v="24"`, 'Sec-Ch-Ua-Mobile': '?0', 'Sec-Ch-Ua-Platform': '"Windows"', 'Upgrade-Insecure-Requests': '1', ...extra, }; } export function buildApiHeaders(cookie: string, extra?: Record): Record { return { 'Cookie': cookie, 'User-Agent': `Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/${CHROME_VERSION}.0.0.0 Safari/537.36`, 'Accept': 'application/json, text/plain, */*', 'Accept-Language': 'zh-CN,zh;q=0.9,en;q=0.8', 'Sec-Fetch-Dest': 'empty', 'Sec-Fetch-Mode': 'cors', 'Sec-Fetch-Site': 'same-origin', 'Sec-Ch-Ua': `"Chromium";v="${CHROME_VERSION}", "Not_A Brand";v="24"`, 'Sec-Ch-Ua-Mobile': '?0', 'Sec-Ch-Ua-Platform': '"Windows"', ...extra, }; } export async function fetchWithRetry( url: string, options: RequestInit, maxRetries = 3, ): Promise { let lastError: Error | null = null; for (let attempt = 0; attempt < maxRetries; attempt++) { if (attempt > 0) { const delay = Math.min(1000 * Math.pow(2, attempt) + Math.random() * 500, 8000); await new Promise(r => setTimeout(r, delay)); } try { const response = await fetch(url, options); if (response.status === 403 && attempt < maxRetries - 1) { console.error(`[fetchWithRetry] 403 on attempt ${attempt + 1}, retrying...`); continue; } return response; } catch (err) { lastError = err as Error; console.error(`[fetchWithRetry] Attempt ${attempt + 1} failed: ${lastError.message}`); } } throw lastError ?? new Error('All retry attempts failed'); } // ─── Cookie Parsing ───────────────────────────────────────── export function parseCookieValue(raw: string, name: string): string | undefined { const match = raw.match(new RegExp(`(?:^|;\\s*)${name}=([^;]*)`)); return match?.[1]; } // ─── Utility types ────────────────────────────────────────── export type ReplyMode = 'approve' | 'auto'; export interface PerformanceStats { total_replies: number; by_platform: Record; by_date: Record; } FILE:scripts/adapters/base.ts // SECURITY MANIFEST: // Environment variables accessed: none // External endpoints called: none // Local files read: none // Local files written: none import type { PlatformAdapter } from '../types.js'; const adapterRegistry: Record Promise> = { 'bilibili': async () => { const { BilibiliAdapter } = await import('./bilibili.js'); return new BilibiliAdapter(); }, 'tieba': async () => { const { TiebaAdapter } = await import('./tieba.js'); return new TiebaAdapter(); }, 'zhihu': async () => { const { ZhihuAdapter } = await import('./zhihu.js'); return new ZhihuAdapter(); }, 'xiaohongshu': async () => { const { XiaohongshuAdapter } = await import('./xiaohongshu.js'); return new XiaohongshuAdapter(); }, }; export async function getAdapter(platformId: string): Promise { const factory = adapterRegistry[platformId]; if (!factory) { const available = Object.keys(adapterRegistry).join(', '); throw new Error(`Unknown platform: "${platformId}". Available: ${available}`); } return factory(); } export function listPlatforms(): string[] { return Object.keys(adapterRegistry); } const isMainModule = process.argv[1]?.replace(/\\/g, '/').endsWith('adapters/base.ts'); if (isMainModule && process.argv[2] === 'test') { console.log(JSON.stringify({ script: 'adapters/base', status: 'ok', platforms: listPlatforms(), })); } FILE:scripts/adapters/bilibili.ts // SECURITY MANIFEST: // Environment variables accessed: none // External endpoints called: https://api.bilibili.com // Local files read: none // Local files written: none import type { PlatformAdapter, Credential, Post, ReplyResult, CheckResult, RateLimitInfo, } from '../types.js'; import { parseCookieValue } from '../types.js'; const BILIBILI_API = 'https://api.bilibili.com'; function buildHeaders(credential: Credential): Record { return { 'Cookie': credential.value, 'User-Agent': 'Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/120.0.0.0 Safari/537.36', 'Referer': 'https://www.bilibili.com', }; } export class BilibiliAdapter implements PlatformAdapter { readonly platformId = 'bilibili'; readonly platformName = '哔哩哔哩'; async search( keyword: string, _timeRange: string, credential: Credential, _target?: string, ): Promise { try { const params = new URLSearchParams({ keyword, search_type: 'video', page: '1', page_size: '20', }); const url = `${BILIBILI_API}/x/web-interface/wbi/search/type?${params}`; const response = await fetch(url, { headers: buildHeaders(credential) }); if (!response.ok) { console.error(`[bilibili] Search failed: ${response.status}`); return []; } const data = await response.json() as { code: number; data?: { result?: Array<{ aid: number; bvid: string; title: string; description: string; author: string; pubdate: number; typeid: number; typename: string; }>; }; }; if (data.code !== 0 || !data.data?.result) { console.error(`[bilibili] Search API error: code=${data.code}`); return []; } return data.data.result.map(item => ({ id: String(item.aid), url: `https://www.bilibili.com/video/${item.bvid}`, title: item.title.replace(/<[^>]*>/g, ''), body: item.description, author: item.author, createdAt: new Date(item.pubdate * 1000).toISOString(), platform: this.platformId, community: item.typename, metadata: { bvid: item.bvid, aid: item.aid }, })); } catch (error) { console.error(`[bilibili] Search error: ${(error as Error).message}`); return []; } } async reply( postId: string, content: string, credential: Credential, ): Promise { try { const csrf = parseCookieValue(credential.value, 'bili_jct'); if (!csrf) { return { success: false, error: 'Missing bili_jct in cookie (required for CSRF)', mode: 'api' }; } const response = await fetch(`${BILIBILI_API}/x/v2/reply/add`, { method: 'POST', headers: { ...buildHeaders(credential), 'Content-Type': 'application/x-www-form-urlencoded', }, body: new URLSearchParams({ type: '1', oid: postId, message: content, csrf, }).toString(), }); if (!response.ok) { return { success: false, error: `HTTP ${response.status}`, mode: 'api' }; } const data = await response.json() as { code: number; message?: string; data?: { rpid?: number }; }; if (data.code !== 0) { return { success: false, error: `B站 API: ${data.message ?? `code ${data.code}`}`, mode: 'api' }; } return { success: true, replyId: data.data?.rpid ? String(data.data.rpid) : undefined, mode: 'api', }; } catch (error) { return { success: false, error: (error as Error).message, mode: 'api' }; } } async check(credential: Credential): Promise { try { const response = await fetch(`${BILIBILI_API}/x/web-interface/nav`, { headers: buildHeaders(credential), }); if (!response.ok) return { valid: false, error: `HTTP ${response.status}` }; const data = await response.json() as { code: number; data?: { isLogin: boolean; uname?: string }; }; if (data.code !== 0 || !data.data?.isLogin) { return { valid: false, error: 'Not logged in or cookie expired' }; } return { valid: true, username: data.data.uname }; } catch (error) { return { valid: false, error: (error as Error).message }; } } rateLimitInfo(): RateLimitInfo { return { requestsPerMinute: 30, repliesPerDay: 30, minReplyIntervalSeconds: 60, notes: 'B站: 搜索≤30次/时, 评论≤50条/天, 重复内容自动过滤', }; } } const isMainModule = process.argv[1]?.replace(/\\/g, '/').endsWith('bilibili.ts'); if (isMainModule && process.argv[2] === 'test') { const adapter = new BilibiliAdapter(); console.log(JSON.stringify({ adapter: 'bilibili', status: 'ok', platformId: adapter.platformId, platformName: adapter.platformName, rateLimit: adapter.rateLimitInfo(), })); } FILE:scripts/adapters/tieba.ts // SECURITY MANIFEST: // Environment variables accessed: none // External endpoints called: https://tieba.baidu.com // Local files read: none // Local files written: none import type { PlatformAdapter, Credential, Post, ReplyResult, CheckResult, RateLimitInfo, BrowserInstruction, } from '../types.js'; import { buildBrowserHeaders, buildApiHeaders, fetchWithRetry } from '../types.js'; const TIEBA_BASE = 'https://tieba.baidu.com'; function pageHeaders(credential: Credential): Record { return buildBrowserHeaders(credential.value, { 'Referer': 'https://tieba.baidu.com/', }); } function apiHeaders(credential: Credential): Record { return buildApiHeaders(credential.value, { 'Referer': 'https://tieba.baidu.com/', 'Origin': 'https://tieba.baidu.com', }); } async function getTbs(credential: Credential): Promise { try { const response = await fetchWithRetry( `${TIEBA_BASE}/dc/common/tbs`, { headers: apiHeaders(credential) }, 2, ); if (!response.ok) return null; const data = await response.json() as { tbs?: string; is_login?: number }; return data.tbs ?? null; } catch { return null; } } export class TiebaAdapter implements PlatformAdapter { readonly platformId = 'tieba'; readonly platformName = '百度贴吧'; async search( keyword: string, _timeRange: string, credential: Credential, target?: string, ): Promise { try { const url = target ? `${TIEBA_BASE}/f?kw=${encodeURIComponent(target)}&ie=utf-8&pn=0` : `${TIEBA_BASE}/f/search/res?qw=${encodeURIComponent(keyword)}&rn=20&pn=1`; const response = await fetchWithRetry(url, { headers: pageHeaders(credential) }); if (response.status === 403) { console.error( `[tieba] Search blocked (403). Use browserSearch() for browser-based fallback.`, ); return []; } if (!response.ok) { console.error(`[tieba] Search failed: ${response.status}`); return []; } const html = await response.text(); if (html.includes('百度安全验证') || html.includes('wappass.baidu.com')) { console.error('[tieba] Hit anti-bot verification page. Use browserSearch() fallback.'); return []; } return this.parseSearchHtml(html, target); } catch (error) { console.error(`[tieba] Search error: ${(error as Error).message}`); return []; } } browserSearch(keyword: string, target?: string): BrowserInstruction { const url = target ? `${TIEBA_BASE}/f?kw=${encodeURIComponent(target)}&ie=utf-8&pn=0` : `${TIEBA_BASE}/f/search/res?qw=${encodeURIComponent(keyword)}&rn=20&pn=1`; return { mode: 'browser', action: 'search', steps: [ { action: 'navigate', url }, { action: 'wait', selector: target ? '#thread_list' : '.s_post' }, { action: 'extract', selector: target ? '#thread_list .j_thread_list' : '.s_post', fields: ['href', 'title', 'text'], }, ], cookies: undefined, }; } private parseSearchHtml(html: string, target?: string): Post[] { const posts: Post[] = []; const threadPattern = /href="\/p\/(\d+)"[^>]*>([^<]+)<\/a>/g; let match; while ((match = threadPattern.exec(html)) !== null) { const [, threadId, title] = match; if (!threadId || !title) continue; const cleanTitle = title.trim(); if (cleanTitle.length < 4) continue; posts.push({ id: threadId, url: `${TIEBA_BASE}/p/${threadId}`, title: cleanTitle, body: '', author: '', createdAt: new Date().toISOString(), platform: this.platformId, community: target, }); } const uniquePosts = new Map(); for (const p of posts) { if (!uniquePosts.has(p.id)) uniquePosts.set(p.id, p); } return Array.from(uniquePosts.values()).slice(0, 20); } async reply( postId: string, content: string, credential: Credential, ): Promise { try { const tbs = await getTbs(credential); if (!tbs) { return { success: false, error: 'Failed to get tbs token', mode: 'api' }; } const metadata = (this as unknown as { _lastReplyMeta?: { kw: string; fid: string } })._lastReplyMeta; const kw = metadata?.kw ?? ''; const fid = metadata?.fid ?? ''; const response = await fetchWithRetry( `${TIEBA_BASE}/f/commit/post/add`, { method: 'POST', headers: { ...apiHeaders(credential), 'Content-Type': 'application/x-www-form-urlencoded', }, body: new URLSearchParams({ ie: 'utf-8', kw, fid, tid: postId, content, tbs, }).toString(), }, 2, ); if (!response.ok) { return { success: false, error: `HTTP ${response.status}`, mode: 'api' }; } const data = await response.json() as { err_code?: number; error?: string; data?: { tid?: string }; }; if (data.err_code && data.err_code !== 0) { return { success: false, error: data.error ?? `err_code ${data.err_code}`, mode: 'api' }; } return { success: true, mode: 'api' }; } catch (error) { return { success: false, error: (error as Error).message, mode: 'api' }; } } async check(credential: Credential): Promise { try { const response = await fetchWithRetry( `${TIEBA_BASE}/dc/common/tbs`, { headers: apiHeaders(credential) }, 2, ); if (!response.ok) return { valid: false, error: `HTTP ${response.status}` }; const data = await response.json() as { tbs?: string; is_login?: number }; if (data.is_login !== 1) { return { valid: false, error: 'Not logged in or BDUSS expired' }; } return { valid: true, username: '(logged in via BDUSS)' }; } catch (error) { return { valid: false, error: (error as Error).message }; } } rateLimitInfo(): RateLimitInfo { return { requestsPerMinute: 20, repliesPerDay: 20, minReplyIntervalSeconds: 120, notes: '贴吧: BDUSS有效期6个月+, 回帖≤30条/天, 重复内容被删, 需从请求头获取BDUSS(非Cookie-Editor的BDUSS_BFESS). 搜索被403时使用browserSearch()回退.', }; } } const isMainModule = process.argv[1]?.replace(/\\/g, '/').endsWith('tieba.ts'); if (isMainModule && process.argv[2] === 'test') { const adapter = new TiebaAdapter(); console.log(JSON.stringify({ adapter: 'tieba', status: 'ok', platformId: adapter.platformId, platformName: adapter.platformName, rateLimit: adapter.rateLimitInfo(), hasBrowserSearch: typeof adapter.browserSearch === 'function', })); } FILE:scripts/adapters/xiaohongshu.ts // SECURITY MANIFEST: // Environment variables accessed: none // External endpoints called: https://edith.xiaohongshu.com/api/sns/web/v1 // Local files read: none // Local files written: none import type { PlatformAdapter, Credential, Post, ReplyResult, CheckResult, RateLimitInfo, BrowserInstruction, } from '../types.js'; import { BROWSER_INSTRUCTION_ID } from '../types.js'; const XHS_API = 'https://edith.xiaohongshu.com/api/sns/web/v1'; function buildHeaders(credential: Credential): Record { return { 'Cookie': credential.value, 'User-Agent': 'Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/120.0.0.0 Safari/537.36', 'Referer': 'https://www.xiaohongshu.com', 'Origin': 'https://www.xiaohongshu.com', }; } export class XiaohongshuAdapter implements PlatformAdapter { readonly platformId = 'xiaohongshu'; readonly platformName = '小红书'; async search( keyword: string, _timeRange: string, credential: Credential, _target?: string, ): Promise { try { const params = new URLSearchParams({ keyword, page: '1', page_size: '20', search_id: '', sort: 'general', note_type: '0', }); const response = await fetch(`${XHS_API}/search/notes?${params}`, { headers: buildHeaders(credential), }); if (response.ok) { const data = await response.json() as { code?: number; success?: boolean; data?: { items?: Array<{ id: string; note_card?: { display_title?: string; desc?: string; user?: { nickname?: string }; time?: number; }; }>; }; }; if (data.success && data.data?.items?.length) { return data.data.items.map(item => ({ id: item.id, url: `https://www.xiaohongshu.com/explore/${item.id}`, title: item.note_card?.display_title ?? '', body: item.note_card?.desc ?? '', author: item.note_card?.user?.nickname ?? '', createdAt: item.note_card?.time ? new Date(item.note_card.time * 1000).toISOString() : new Date().toISOString(), platform: this.platformId, })); } } } catch (error) { console.error(`[xiaohongshu] API search failed, falling back to browser: ${(error as Error).message}`); } console.error('[xiaohongshu] API search unavailable, returning browser instruction'); const instruction: BrowserInstruction = { mode: 'browser', action: 'search', steps: [ { action: 'navigate', url: `https://www.xiaohongshu.com/search_result?keyword=${encodeURIComponent(keyword)}&source=web_search_result_notes` }, { action: 'wait', selector: '.note-item' }, { action: 'extract', selector: '.note-item', fields: ['title', 'url', 'author', 'likes'] }, ], cookies: credential.value, }; return [{ id: BROWSER_INSTRUCTION_ID, url: '', title: '', body: JSON.stringify(instruction), author: '', createdAt: new Date().toISOString(), platform: this.platformId, }]; } async reply( postId: string, content: string, credential: Credential, ): Promise { const instruction: BrowserInstruction = { mode: 'browser', action: 'reply', steps: [ { action: 'navigate', url: `https://www.xiaohongshu.com/explore/${postId}` }, { action: 'wait', selector: '#content-textarea' }, { action: 'click', selector: '#content-textarea' }, { action: 'type', text: content }, { action: 'click', selector: '.submit-btn' }, ], cookies: credential.value, }; return { success: true, replyId: `__browser_pending__:${JSON.stringify(instruction)}`, mode: 'browser', }; } async check(credential: Credential): Promise { if (!credential.value || credential.value.length === 0) { return { valid: false, error: 'No cookie provided' }; } return { valid: true, username: '(cookie mode — verify via browser, cookies expire ~12h)', }; } rateLimitInfo(): RateLimitInfo { return { requestsPerMinute: 20, repliesPerDay: 10, minReplyIntervalSeconds: 3, notes: 'No public API. Browser-only. Cookies expire ~12h. Min 3s between requests. Strongly recommend SocialVault for auto cookie refresh.', }; } } const isMainModule = process.argv[1]?.replace(/\\/g, '/').endsWith('xiaohongshu.ts'); if (isMainModule && process.argv[2] === 'test') { const adapter = new XiaohongshuAdapter(); console.log(JSON.stringify({ adapter: 'xiaohongshu', status: 'ok', platformId: adapter.platformId, platformName: adapter.platformName, rateLimit: adapter.rateLimitInfo(), })); } FILE:scripts/adapters/zhihu.ts // SECURITY MANIFEST: // Environment variables accessed: none // External endpoints called: https://www.zhihu.com/api/v4 // Local files read: none // Local files written: none import type { PlatformAdapter, Credential, Post, ReplyResult, CheckResult, RateLimitInfo, BrowserInstruction, } from '../types.js'; import { BROWSER_INSTRUCTION_ID, buildApiHeaders, buildBrowserHeaders, fetchWithRetry } from '../types.js'; const ZHIHU_API = 'https://www.zhihu.com/api/v4'; function searchApiHeaders(credential: Credential): Record { return buildApiHeaders(credential.value, { 'Referer': 'https://www.zhihu.com/search', 'Origin': 'https://www.zhihu.com', 'x-requested-with': 'fetch', }); } function checkApiHeaders(credential: Credential): Record { return buildApiHeaders(credential.value, { 'Referer': 'https://www.zhihu.com/', 'Origin': 'https://www.zhihu.com', }); } export class ZhihuAdapter implements PlatformAdapter { readonly platformId = 'zhihu'; readonly platformName = '知乎'; async search( keyword: string, _timeRange: string, credential: Credential, _target?: string, ): Promise { try { const params = new URLSearchParams({ gk_version: 'gz-gaokao', t: 'general', q: keyword, correction: '1', offset: '0', limit: '20', filter_fields: '', lc_idx: '0', show_all_topics: '0', search_source: 'Normal', }); const response = await fetchWithRetry( `${ZHIHU_API}/search_v3?${params}`, { headers: searchApiHeaders(credential) }, ); if (response.status === 403) { console.error( '[zhihu] Search blocked (403). Likely missing x-zse-96 signature. Use browserSearch() fallback.', ); return []; } if (!response.ok) { console.error(`[zhihu] Search failed: ${response.status}`); return []; } const data = await response.json() as { data?: Array<{ type?: string; object?: { id?: number; type?: string; question?: { id?: number; title?: string }; title?: string; excerpt?: string; content?: string; author?: { name?: string }; created_time?: number; updated_time?: number; }; }>; }; if (!data.data) return []; const posts: Post[] = []; for (const item of data.data) { const obj = item.object; if (!obj) continue; if (item.type === 'search_result' && obj.question) { posts.push({ id: String(obj.question.id ?? obj.id ?? ''), url: `https://www.zhihu.com/question/${obj.question.id}`, title: obj.question.title ?? '', body: (obj.excerpt ?? obj.content ?? '').replace(/<[^>]*>/g, ''), author: obj.author?.name ?? '', createdAt: obj.created_time ? new Date(obj.created_time * 1000).toISOString() : new Date().toISOString(), platform: this.platformId, }); } } const unique = new Map(); for (const p of posts) { if (p.id && !unique.has(p.id)) unique.set(p.id, p); } return Array.from(unique.values()); } catch (error) { console.error(`[zhihu] Search error: ${(error as Error).message}`); return []; } } browserSearch(keyword: string, _target?: string): BrowserInstruction { const url = `https://www.zhihu.com/search?type=content&q=${encodeURIComponent(keyword)}`; return { mode: 'browser', action: 'search', steps: [ { action: 'navigate', url }, { action: 'wait', selector: '.SearchResult-Card' }, { action: 'extract', selector: '.SearchResult-Card', fields: ['href', 'title', 'text'], }, ], cookies: undefined, }; } async reply( postId: string, content: string, credential: Credential, ): Promise { const instruction: BrowserInstruction = { mode: 'browser', action: 'reply', steps: [ { action: 'navigate', url: `https://www.zhihu.com/question/${postId}` }, { action: 'wait', selector: '.AnswerForm' }, { action: 'click', selector: '.AnswerForm .RichContent' }, { action: 'type', text: content }, { action: 'click', selector: 'button[type="submit"]' }, ], cookies: credential.value, }; return { success: true, replyId: `__browser_pending__:${JSON.stringify(instruction)}`, mode: 'browser', }; } async check(credential: Credential): Promise { try { const response = await fetchWithRetry( `${ZHIHU_API}/me`, { headers: checkApiHeaders(credential) }, 2, ); if (!response.ok) return { valid: false, error: `HTTP ${response.status}` }; const data = await response.json() as { id?: string; name?: string; error?: { message?: string } }; if (data.error) { return { valid: false, error: data.error.message ?? 'Unknown error' }; } return { valid: !!data.id, username: data.name }; } catch (error) { return { valid: false, error: (error as Error).message }; } } rateLimitInfo(): RateLimitInfo { return { requestsPerMinute: 10, repliesPerDay: 10, minReplyIntervalSeconds: 300, notes: '知乎: z_c0有效期~30天, 反爬严格(需x-zse-96签名), 高频触发验证码. 写入用browser模式. 搜索被403时使用browserSearch()回退.', }; } } const isMainModule = process.argv[1]?.replace(/\\/g, '/').endsWith('zhihu.ts'); if (isMainModule && process.argv[2] === 'test') { const adapter = new ZhihuAdapter(); console.log(JSON.stringify({ adapter: 'zhihu', status: 'ok', platformId: adapter.platformId, platformName: adapter.platformName, rateLimit: adapter.rateLimitInfo(), hasBrowserSearch: typeof adapter.browserSearch === 'function', })); } FILE:scripts/adapters/_template.ts // SECURITY MANIFEST: // Environment variables accessed: (list here) // External endpoints called: (list here) // Local files read: (list here) // Local files written: (list here) import type { PlatformAdapter, Credential, Post, ReplyResult, CheckResult, RateLimitInfo, } from '../types.js'; /** * Template for creating a new platform adapter. * * Steps: * 1. Copy this file to scripts/adapters/.ts * 2. Rename the class and update platformId / platformName * 3. Implement search(), reply(), check(), rateLimitInfo() * 4. Register in scripts/adapters/base.ts * 5. Create templates/reply-.md * 6. Add rate limits to references/safety-rules.md * 7. Add TOS notes to references/platform-tos-notes.md * 8. Test with: npx tsx scripts/adapters/.ts test */ export class TemplateAdapter implements PlatformAdapter { readonly platformId = 'template'; readonly platformName = 'Template Platform'; async search( keyword: string, timeRange: string, credential: Credential, target?: string, ): Promise { // TODO: Implement platform-specific search // API mode: use fetch() with credential // Browser mode: return BrowserInstruction wrapped in Post[] console.error(`[${this.platformId}] search() not implemented`); return []; } async reply( postId: string, content: string, credential: Credential, ): Promise { // TODO: Implement reply logic console.error(`[${this.platformId}] reply() not implemented`); return { success: false, error: 'Not implemented' }; } async check(credential: Credential): Promise { // TODO: Implement credential validation console.error(`[${this.platformId}] check() not implemented`); return { valid: false, error: 'Not implemented' }; } rateLimitInfo(): RateLimitInfo { return { requestsPerMinute: 60, repliesPerDay: 10, minReplyIntervalSeconds: 300, notes: 'TODO: Add platform-specific rate limit notes', }; } } const isMainModule = process.argv[1]?.replace(/\\/g, '/').endsWith('_template.ts'); if (isMainModule && process.argv[2] === 'test') { const adapter = new TemplateAdapter(); console.log(JSON.stringify({ adapter: 'template', status: 'ok', platformId: adapter.platformId, platformName: adapter.platformName, rateLimit: adapter.rateLimitInfo(), })); } FILE:references/platform-tos-notes.md # 平台服务条款关键摘要 ## B站 - 搜索接口: `api.bilibili.com/x/web-interface/wbi/search` - 评论接口: `api.bilibili.com/x/v2/reply/add`（需 CSRF=bili_jct） - 搜索频率建议 ≤30 次/小时 - 评论频率建议 ≤50 条/天 - 重复内容会被自动过滤 - 对服务器 IP 无特殊限制 ## 百度贴吧 - BDUSS 有效期 6 个月+，非常稳定 - Cookie-Editor 导出的是 BDUSS_BFESS（不可用），必须从网络请求头获取 BDUSS - 回帖需先获取 tbs token - 回帖频率建议 ≤30 条/天 - 重复内容会被系统删除 - STOKEN 用于写操作验证 ## 知乎 - 搜索接口: `zhihu.com/api/v4/search_v3` - 回答需要 x-zse-96 签名（当前用 browser 模式绕过） - z_c0 有效期约 30 天 - 对服务器 IP 有限制，高频访问触发验证码 - 回答/评论有内容审核 ## 小红书 - 无公开写入 API，回复操作通过浏览器模拟 - 搜索接口: `edith.xiaohongshu.com/api/sns/web/v1/search/notes`（需 X-s/X-t 签名） - Cookie 有效期极短: ~12 小时 - 多层反爬: Header + 行为指纹 + 设备ID + WASM - 强烈建议配合 SocialVault 自动续期 - 评论区不允许外部链接 - 请求间隔不低于 3 秒 FILE:references/safety-rules.md # SeedDrop 安全规则这些规则在脚本中硬编码执行，不可通过配置绕过。 ## 频率限制 ### 每平台每日回复上限 | 平台 | Approve 模式 | Auto 模式 | 平台限制参考 | |------|-------------|-----------|-------------| | B站 | 30 | 15 | 评论 ≤50/天, 搜索 ≤30/时 | | 贴吧 | 20 | 10 | 回帖 ≤30/天, 重复内容删除 | | 知乎 | 10 | 5 | 反爬严格, 高频触发验证码 | | 小红书 | 10 | 5 | 无 API, ≤1 req/3s | | 其他平台 | 10 | 5 | — | Auto 模式的上限为 Approve 模式的 50%，作为额外安全措施。 ### 回复间隔 - 最小间隔: 5 分钟 - 最大间隔: 15 分钟 - 实际间隔: 在 5-15 分钟范围内随机（不固定） - 小红书额外要求: 任意两次请求间隔不低于 3 秒 ### 同一作者限制 - 同一用户的帖子 24 小时内最多回复 1 次 - 按作者 ID + 平台维度去重 ### 同一帖子限制 - 每个帖子只能回复 1 次 - 通过 `interaction-log.jsonl` 中的 `post_id` 去重 ## 内容安全 - 每条回复必须提供实质性帮助 - 品牌提及不超过回复总内容的 20% - 不使用营销话术、夸大词汇 - 不包含直接联系方式或裸链接 ## 回复多样性 - 不使用固定模板（每次重新生成） - 随机化开头方式、长度、是否提及品牌 ## 禁止回复的场景 - 帖子标记为置顶、官方、管理员帖 - 社区规则明确禁止自我推广 - 帖子作者在黑名单中 - 帖子内容涉及敏感话题 ## Auto 模式额外限制 1. 每日上限降至 Approve 模式的 50% 2. 评分阈值最低不能低于 0.7 3. 风险评分低于 0.5 的帖子不回复 4. 新添加的平台默认 7 天内只能用 Approve 模式 ## 黑名单管理存储在 `memory/blacklist.md`，支持用户/社区/关键词三种维度。 ## 法律合规 - 本工具仅辅助用户进行正常社区互动，不是数据采集/爬虫工具 - 使用用户自己的账号凭证操作，不构成未授权访问 - 不大规模采集、存储或传播第三方用户数据 - 内置频率限制确保不干扰平台正常运营 - 用户需自行遵守相关法律法规及各平台服务条款 - 请勿修改或绕过上述安全限制 FILE:references/scoring-criteria.md # SeedDrop 评分标准 ## 评分维度 | 维度 | 权重 | 说明 | |------|------|------| | 相关度 (Relevance) | 35% | 帖子内容与品牌关键词的匹配程度 | | 意图强度 (Intent) | 30% | 帖子作者是否在寻求帮助或建议 | | 时效性 (Freshness) | 20% | 帖子发布时间越新越好 | | 风险评估 (Risk) | 15% | 回复该帖子的潜在风险 | ## 总分计算 ``` final = relevance × 0.35 + intent × 0.30 + freshness × 0.20 + risk × 0.15 ``` ## 阈值 - 默认: 0.6 - Auto 模式最低: 0.7 - 推荐范围: 0.4 - 0.8 ## 相关度评分 | 命中比例 | 分数 | |---------|------| | > 50% 关键词命中 | 1.0 | | 20-50% | 0.7 | | > 0% | 0.5 | | 0% | 0.2 | ## 意图评分 | 类型 | 分数 | 特征词 | |------|------|--------| | 求助/提问 | 0.9 | 求推荐, 怎么, 请问, 有没有推荐, recommend, help | | 讨论/意见 | 0.7 | 讨论, 大家觉得, 有没有人用过, 什么体验 | | 吐槽/抱怨 | 0.5 | 坑, 吐槽, 垃圾, 难用, 避雷 | | 纯分享 | 0.3 | 无特征词 | ## 时效评分 | 帖子年龄 | 分数 | |---------|------| | ≤ 2h | 1.0 | | 2-6h | 0.9 | | 6-12h | 0.8 | | 12-24h | 0.6 | | 24-48h | 0.4 | | > 48h | 0.2 | ## 风险评分 | 类型 | 分数 | |------|------| | 安全帖子 | 0.9 | | 官方/管理相关 | 0.3 | | 争议性话题 | 0.2 | FILE:memory/blacklist.md # SeedDrop 黑名单 ## 用户黑名单 (不回复以下用户的帖子) ## 社区黑名单 (不在以下社区/subreddit 发布) ## 关键词黑名单 (帖子包含以下关键词时不回复) - [Meta] - [Announcement] - [Mod Post] - [Official] FILE:memory/brand-profile.md # SeedDrop Brand Profile ## 基本信息 - 业务名称: SeedDrop - 业务类型: 社区互动自动化工具 - 服务区域: 中文互联网（B站、贴吧、知乎、小红书） - 官网: (待配置) - 核心卖点: 自动监控社区讨论，AI生成高质量回复，自然种草，帮助独立开发者和小团队低成本做内容运营 ## 目标客户关键词 - 主关键词: 社区营销, 内容运营工具, 自动回复, 种草工具 - 场景关键词: 独立开发运营, SaaS推广, 小团队营销, 社交媒体运营效率 - 竞品关键词: 社媒管理工具, 内容营销平台 ## 品牌人设 - 语气: 真诚分享、像开发者朋友一样聊天，不硬推销 - 禁用语: 最好的, 第一, 保证, 神器, 爆款, 裂变, 私域 - 常用语: 我们也在用这个方案, 刚好做了个小工具, 分享一下踩过的坑 - 可提及案例: 用 SeedDrop 自动监控「独立开发」「SaaS工具」等话题，每天节省2小时人工筛帖时间 ## 关键词 - 社区营销 - 内容运营工具 - 自动回复 - 种草助手 - 独立开发 - 内容营销 - 运营效率 - SaaS工具推广 - 社交媒体运营 - 社区互动 ## 平台 - bilibili - tieba - zhihu - xiaohongshu ## 监控平台配置 platforms: - id: bilibili enabled: true keywords: [] - id: tieba enabled: true bars: [] keywords: [] - id: zhihu enabled: true keywords: [] - id: xiaohongshu enabled: true keywords: [] ## 运行模式 - mode: approve - scoring_threshold: 0.6 - daily_max_replies: bilibili: 30 tieba: 20 zhihu: 10 xiaohongshu: 10 ## 语言偏好 - primary: zh-CN - secondary: en FILE:guides/adapter-development.md # 平台适配器开发指南 ## 开发步骤 ### 1. 复制模板 ```bash cp scripts/adapters/_template.ts scripts/adapters/.ts ``` ### 2. 实现 PlatformAdapter 接口 ```typescript import type { PlatformAdapter, Credential, Post, ReplyResult, CheckResult, RateLimitInfo } from '../types.js'; export class MyAdapter implements PlatformAdapter { readonly platformId = 'my-platform'; readonly platformName = 'My Platform'; async search(keyword: string, timeRange: string, credential: Credential, target?: string): Promise { ... } async reply(postId: string, content: string, credential: Credential): Promise { ... } async check(credential: Credential): Promise { ... } rateLimitInfo(): RateLimitInfo { ... } } ``` ### 3. 输出格式 **search 输出**（JSONL，每行一个）： ```json {"id":"...","url":"...","title":"...","body":"...","author":"...","createdAt":"...","platform":""} ``` **reply 输出**： ```json {"success": true, "replyId": "..."} ``` **check 输出**： ```json {"valid": true, "username": "..."} ``` ### 4. 两种实现方式 **有 API 的平台**：使用 `fetch()` 直接调用 API，参考 `reddit.ts`。 **无 API 的平台**：返回 browser 指令，参考 `xiaohongshu.ts`： ```typescript import { BROWSER_INSTRUCTION_ID, type BrowserInstruction } from '../types.js'; const instruction: BrowserInstruction = { mode: 'browser', action: 'search', steps: [ { action: 'navigate', url: `https://example.com/search?q=${keyword}` }, { action: 'wait', selector: '.result-item' }, { action: 'extract', selector: '.result-item', fields: ['title', 'url'] }, ], cookies: credential.value, }; return [{ id: BROWSER_INSTRUCTION_ID, url: '', title: '', body: JSON.stringify(instruction), author: '', createdAt: new Date().toISOString(), platform: this.platformId }]; ``` ### 5. 注册适配器在 `scripts/adapters/base.ts` 中添加： ```typescript 'my-platform': async () => { const { MyAdapter } = await import('./my-platform.js'); return new MyAdapter(); }, ``` ### 6. 配套文件 | 文件 | 用途 | |------|------| | `templates/reply-.md` | 回复风格指南 | | `references/safety-rules.md` | 添加频率限制 | | `references/platform-tos-notes.md` | 添加 ToS 摘要 | ### 7. 测试 ```bash npx tsx scripts/adapters/.ts test ``` ### 8. 启用在 `memory/brand-profile.md` 的 platforms 中添加新平台。 FILE:guides/brand-profile-setup.md # Brand Profile 配置引导 ## 配置流程执行 `seeddrop setup` 时按以下步骤引导： ### 第 1 步：基本信息 - 业务名称、类型、服务区域 - 官网链接（不主动放入回复） - 核心卖点（1-3 句） ### 第 2 步：关键词 - 主关键词（3-10 个） - 场景关键词（3-10 个） - 竞品关键词（可选） - 中英文都要覆盖 ### 第 3 步：品牌人设 - 语气风格：专业亲切 / 轻松幽默 / 正式严谨 / 热情直接 - 禁用语：如"最好的""保证""第一" - 常用语和可引用案例 ### 第 4 步：平台配置 - 选择启用的平台 - Reddit: 目标 subreddit 列表 - X/Twitter: 话题标签 - 小红书: 搜索关键词 ### 第 5 步：运行模式 - 模式: approve（默认）/ auto - 评分阈值: 默认 0.6 - 每日上限: 各平台独立配置 ### 第 6 步：语言偏好 - 主语言 / 辅助语言 ## 生成文件配置完成写入 `memory/brand-profile.md`，格式参照 `docs/brand-profile.md` 示例。 FILE:guides/quickstart.md # SeedDrop 快速上手 ## 1. 安装 ```bash clawhub install seeddrop ``` ## 2. 配置品牌资料 ```bash seeddrop setup ``` 按提示输入： - 产品/品牌名称 - 一句话描述 - 监控关键词（多个用逗号分隔） - 使用的平台（bilibili, tieba, zhihu, xiaohongshu） - 回复模式（approve = 人工审核, auto = 自动发送） ## 3. 添加平台凭证 ```bash seeddrop auth add bilibili seeddrop auth add tieba ``` 需要提供对应平台的 Cookie。推荐配合 SocialVault 管理凭证： ```bash clawhub install socialvault socialvault add bilibili ``` ## 4. 运行监控 ```bash seeddrop monitor bilibili seeddrop monitor tieba 编程 seeddrop monitor all ``` ## 5. 查看报告 ```bash seeddrop report seeddrop report weekly ``` ## 平台 Cookie 获取 | 平台 | 必需字段 | 获取方式 | |------|---------|---------| | B站 | SESSDATA, bili_jct | 浏览器 DevTools → Application → Cookies | | 贴吧 | BDUSS, STOKEN | 浏览器 DevTools → Network → 请求头 Cookie | | 知乎 | z_c0, d_c0 | 浏览器 DevTools → Application → Cookies | | 小红书 | a1, web_session | 浏览器 DevTools → Application → Cookies | > **注意**：贴吧的 BDUSS 必须从请求头获取，Cookie-Editor 导出的 BDUSS_BFESS 不可用。 ## 推荐搭配安装 **SocialVault** 后可自动管理 Cookie 续期，特别是小红书（Cookie 仅 12 小时有效）。 FILE:.qoder/rules/code-quality.md --- trigger: always_on --- # 代码质量与测试标准 ## 命名规范 - 变量名、函数名必须自解释，禁止 temp、data、info、handle 等模糊命名 - 布尔变量用 is/has/should/can 前缀 - 函数用动词开头：calculateTotalPrice 而非 totalPrice - 常量用 UPPER_SNAKE_CASE - 遵循项目已有命名约定（先 grep 确认） ## 函数设计 - 单一职责：一个函数只做一件事 - 函数体不超过 40 行，超过则拆分 - 参数不超过 4 个，超过则用 Options Object - 禁止布尔参数控制分支（用策略模式或独立函数） - 强类型语言中每个函数必须有明确返回值类型 ## 错误处理 - 禁止空 catch 块，每个 catch 必须有日志或重抛 - 外部 I/O 必须有错误处理 - 使用有意义的自定义错误类型 - 异步代码中所有 Promise 必须有错误处理链 - 用户可见错误信息必须友好且不泄漏内部细节 ## 代码结构 - 导入分组：外部库 → 内部模块 → 类型 → 样式 - 相关逻辑物理邻近，不相关逻辑物理分离 - 禁止超过 3 层嵌套（用 early return 或函数抽取） - 魔法数字和魔法字符串必须提取为命名常量 ## 注释哲学 - 好代码本身就是注释，优先靠命名和结构自解释 - 只在需要解释"为什么"时写注释，不解释"是什么" - 公共 API 必须有 JSDoc / docstring - TODO 必须附 issue 链接或负责人 ## 测试标准 ### 验证闭环每次代码变更后必须执行： 1. get_problems → 修到 0 error 2. 运行受影响模块测试 → 全部通过 3. 涉及模块交互时运行集成测试 4. 逐条核对验收标准任何一环不通过则自主修复，禁止标记完成 ### 测试文件生成 - 一次只生成一个测试文件 - 生成后立即 get_problems 检查编译 - 修复编译问题后执行测试 - 当前文件通过后才进入下一个 ### 测试用例设计 - 每个测试用例独立，不依赖执行顺序 - 命名格式：should_[预期行为]_when_[条件] - 必须覆盖：Happy path / Edge cases / Error scenarios / 回归防护 - 每个测试只断言一个行为 - 测试行为和结果，不测实现细节 - Mock 只用于隔离外部依赖 FILE:.qoder/rules/core-principles.md --- trigger: always_on --- # Vue 2 项目专项规则本规则适用于 Vue 2.x（含 2.7）项目，使用 Options API 风格。 ## 版本识别在执行任何修改前，先确认： - package.json 中 vue 版本为 ^2.x - 构建工具为 Vue CLI (@vue/cli) 或 Webpack - 状态管理为 Vuex（3.x） - UI 库为 Element UI 或其他 Vue 2 兼容库如果发现实际是 Vue 3 项目，立即停止并切换到 Vue 3 规则 ## 组件规范 ### 文件组织 - 单文件组件（SFC）按 template → script → style 顺序排列 - 组件文件名使用 PascalCase：UserProfile.vue - 每个文件只包含一个组件 ### Options API 书写顺序严格按照以下顺序排列 options： 1. name 2. components 3. directives / filters 4. mixins 5. props 6. data 7. computed 8. watch 9. 生命周期钩子（按执行顺序）：beforeCreate → created → beforeMount → mounted → beforeUpdate → updated → beforeDestroy → destroyed 10. methods ### Props 规范 - props 必须有类型定义，禁止纯数组写法 - 正确：props: { title: { type: String, required: true, default: '' } } - 错误：props: ['title'] - 必要时提供 validator 函数 - 复杂对象的 default 必须用工厂函数返回 ### Data 规范 - data 必须是一个返回对象的函数，禁止直接写对象 - data 中不放与模板无关的变量（不需要响应式的数据放在 created 里用 this.xxx = ... 赋值） ### Computed 与 Watch - 能用 computed 解决的不用 watch - watch 中避免复杂逻辑，复杂操作抽到 methods - 需要 deep watch 时显式声明 deep: true 并注释原因 ### 事件规范 - 子组件向父组件通信用 $emit，事件名用 kebab-case - 禁止滥用 $parent / $children 直接访问 - 禁止滥用 EventBus 做跨层级通信（大范围通信走 Vuex） ## Vuex 规范 - State：只存全局共享数据，组件私有状态留在组件内 - Mutations：只做同步状态变更，命名用 SET_XXX / UPDATE_XXX - Actions：处理异步逻辑，命名用动词：fetchUserList / submitOrder - Getters：用于派生状态，等同于 Store 级别的 computed - 模块化：使用 namespaced: true，按业务领域拆分 module ## 路由规范（Vue Router 3.x） - 路由配置使用懒加载：component: () => import('@/views/xxx.vue') - 路由 name 使用 PascalCase，与组件文件名一致 - 路由守卫中的异步操作必须有错误处理 - 禁止在组件内直接操作 window.location，使用 $router ## 样式规范 - 使用 scoped 样式避免全局污染 - 需要穿透子组件时用 ::v-deep 或 /deep/（Vue 2 语法） - 禁止在 scoped style 中使用标签选择器 - BEM 或项目已有的命名约定（先确认再使用） ## 兼容性注意 - Vue 2 不支持 Fragments（模板只能有一个根元素） - Vue 2 不支持 Teleport / Suspense - 数组变更检测限制：使用 Vue.set 或 this.$set 修改数组索引 - 对象新增属性检测限制：使用 Vue.set 或 this.$set 添加新属性 - 如果项目是 Vue 2.7，可使用 Composition API（setup 语法糖除外），但必须先确认项目是否引入了 @vue/composition-api 或已升级到 2.7 ## 禁止事项 - 禁止使用已废弃的 API：$on / $off / $once（Vue 2.x 原生支持但 Vue 3 已移除，迁移风险高） - 禁止在模板中使用复杂的内联表达式（超过一个函数调用的逻辑必须移到 computed 或 methods） - 禁止修改 props（如需修改，用 data 接收或 computed 转换） - 禁止在 data 中引用 props 的默认值后不追踪变化（常见 bug） FILE:.qoder/rules/documentation-gen.md --- trigger: always_on --- # 文档生成规范通过 @documentation-gen 手动触发。 ## README 模板 - 简介：一句话描述项目用途 - 快速开始：最少步骤跑起来 - 架构概览：核心模块及关系 - 开发指南：环境搭建、常用命令 - API 文档（如适用） - 贡献指南（如适用） ## 变更日志 - 使用 Keep a Changelog 格式 - 按 Added / Changed / Fixed / Removed 分类 - 每条关联 issue 或 PR 编号 ## API 文档 - 包含请求方法、路径、参数、响应示例、错误码 - 提供 curl 示例 FILE:.qoder/rules/guardrails.md --- trigger: always_on --- # 防翻车护栏机制 ## 认知护栏 ### 不确定性声明对某个技术细节置信度低于 80% 时，必须标注： "⚠️ 我对此不够确定：[内容]。建议验证方式：[方法]" 禁止用自信语气描述不确定的方案。 ### 假设追踪每个假设必须显式记录： "📌 假设：[内容] | 依据：[来源] | 风险：[假设错误的后果]" 验证阶段逐条回顾假设是否成立。 ### 上下文不足时 - 禁止凭"常见做法"猜测项目特定配置 - 必须用工具验证（grep_code 搜模式、read_file 查配置） - 工具也无法确认时向用户明确询问 ## 执行护栏 ### 变更范围锁定 - 每次改动必须与当前任务直接相关 - "顺便改进"的想法记录到 Memory，当前不动 - 任务范围变化时必须更新 Spec 并获得确认 ### 渐进式修改 - 大型修改分步进行，每步之间运行验证 - 禁止一次性重写整个文件，用 search_replace 精确修改 - 每步改完立即 get_problems，确认无回归再继续 ### 并发安全 - 文件编辑：必须串行 - 终端命令：必须串行 - 只读操作（read_file、search_codebase、grep_code）：可以并行 ### 回滚意识 - 高风险操作前确认有回滚路径 - 可能破坏现有功能的修改先确认测试覆盖 - 数据库 migration、配置变更等不可逆操作必须获得用户明确确认 ## 思维链护栏 ### 防止回归倾向 - 面对复杂任务禁止退缩为"建议用户自行处理" - 遇到阻碍先尝试至少 2 种不同解决路径 - 穷尽工具能力后才向用户求助 ### 防止死循环同一修复尝试 3 次仍失败时必须： 1. 停下来重新分析根因 2. 告知用户当前状态和已尝试方案 3. 提出不同方向的解决思路 ### 防止范围蔓延每完成一步回顾：当前在做的事是否仍在 Spec 范围内？发现偏离则立即停止并重新评估。 ## 输出护栏 ### 代码输出 - 默认用 search_replace 修改文件，不直接输出代码块 - 用户明确要求时才在对话中展示代码 - 展示代码用 // ... existing code ... 省略未改部分 ### 任务完成确认任务完成时必须提供结构化报告： - 变更文件清单 - 验收标准逐条达成情况 - 已运行测试及结果 - 潜在风险提示（如有）未实际完成的任务绝不标记为 complete。 FILE:.qoder/rules/review-and-refactor.md --- trigger: always_on --- # 代码审查与重构 ## 审查清单 ### P0 必须修复 - 逻辑错误：分支遗漏、条件判断错误、循环边界 - 安全漏洞：注入风险、敏感信息泄漏、未授权访问 - 数据一致性：竞态条件、事务缺失、状态不一致 - 资源泄漏：未关闭的连接/文件句柄、内存泄漏 ### P1 强烈建议修复 - 错误处理缺失或不当 - 类型安全问题 - 性能问题（N+1 查询、不必要的重复计算） - 测试覆盖不足 ### P2 改进建议 - 命名可读性 - 代码结构优化 - 注释补充 - 代码重复（DRY） ## 重构原则 - 重构前必须确认测试覆盖 - 一次只做一种类型的重构 - 重构不改变外部行为，每步用测试确认 - 大型重构必须先产出 Spec 获得确认 FILE:.qoder/rules/security-checklist.md --- trigger: always_on --- # 安全检查清单 ## 输入处理 - 所有用户输入必须验证和消毒（sanitize） - SQL 查询使用参数化查询，禁止字符串拼接 - 文件路径操作防范路径穿越（path traversal） - 正则表达式防范 ReDoS ## 认证授权 - 密码和密钥禁止硬编码 - 敏感配置使用环境变量 - API 端点必须有权限校验 - Session/Token 有合理过期策略 ## 数据安全 - 敏感数据不出现在日志中 - API 响应不泄漏内部实现细节 - 错误信息对用户友好且不暴露技术栈 - PII 数据加密存储 ## 依赖安全 - 不引入有已知漏洞的依赖版本 - 优先使用社区活跃维护的库

SeedDrop

Comments (0)