{"id":"845a9852-0566-4984-b974-e08dc1c5e608","entityType":"agent","slug":"clawhub-fandywang87-paper-translation","name":"论文精读翻译","canonicalUrl":"https://www.xpersona.co/agent/clawhub-fandywang87-paper-translation","canonicalPath":"/agent/clawhub-fandywang87-paper-translation","generatedAt":"2026-10-10T21:53:42.675Z","source":"CLAWHUB","claimStatus":"UNCLAIMED","verificationTier":"NONE","summary":{"evidence":{"source":"editorial-content","verified":true,"confidence":"high","updatedAt":"2026-10-10T15:42:59.534Z","emptyReason":null},"description":"ArXiv 论文精读级中文翻译，同步到 IMA 知识库 + 腾讯文档。 基于 5 篇论文（MDL/Kunlun/OneTrans/RankMixer/MixFormer）3 轮迭代实战经验。 触发场景：翻译论文、翻译 arxiv、论文精读、论文中文翻译、paper translation、 translate p... Skill: 论文精读翻译 Owner: fandywang87 Summary: ArXiv 论文精读级中文翻译，同步到 IMA 知识库 + 腾讯文档。 基于 5 篇论文（MDL/Kunlun/OneTrans/RankMixer/MixFormer）3 轮迭代实战经验。 触发场景：翻译论文、翻译 arxiv、论文精读、论文中文翻译、paper translation、 translate p... Tags: latest:1.0.0 Version history: v1.0.0 | 2026-04-21T08:35:01.688Z | user 首次发布：基于 5 篇论文 3 轮迭代的 SOP v2 经验，包含翻译流程、校验脚本、平台兼容性踩坑经验 Archive index: Archive v1.0.0: 6 files, 10030 bytes Files: references/iteration-history","descriptionLabel":"Technical summary","evidenceSummary":"Capability contract not published. No trust telemetry is available yet. 1.4K downloads reported by the source. Last updated 10/10/2026.","installCommand":"clawhub skill install s17czwytgebtz5mxnn62cn9zth855nvb:paper-translation","sourceUrl":"https://clawhub.ai/fandywang87/paper-translation","homepage":"https://clawhub.ai/fandywang87/skills/paper-translation","primaryLinks":[{"label":"View on ClawHub","url":"https://clawhub.ai/fandywang87/paper-translation","kind":"source"},{"label":"Homepage","url":"https://clawhub.ai/fandywang87/skills/paper-translation","kind":"homepage"}],"safetyScore":84,"overallRank":62,"popularityScore":63,"trustScore":null,"claimedByName":null,"isOwner":false,"seoDescription":"ArXiv 论文精读级中文翻译，同步到 IMA 知识库 + 腾讯文档。 基于 5 篇论文（MDL/Kunlun/OneTrans/RankMixer/MixFormer）3 轮迭代实战经验。 触发场景：翻译论文、翻译 arxiv、论文精读、论文中文翻译、paper translation、 translate p..."},"coverage":{"evidence":{"source":"public-profile","verified":false,"confidence":"medium","updatedAt":"2026-10-10T15:42:59.534Z","emptyReason":null},"protocols":[{"protocol":"OPENCLEW","label":"OpenClaw","status":"self-declared","notes":"Declared in the public agent profile."}],"capabilities":[],"verifiedCount":0,"selfDeclaredCount":1,"capabilityMatrix":{"rows":[{"key":"OPENCLEW","type":"protocol","support":"unknown","confidenceSource":"profile","notes":"Listed on profile"}],"flattenedTokens":"protocol:OPENCLEW|unknown|profile"}},"adoption":{"evidence":{"source":"CLAWHUB","verified":false,"confidence":"medium","updatedAt":"2026-10-10T15:42:59.534Z","emptyReason":null},"stars":null,"forks":null,"downloads":1355,"packageName":null,"latestVersion":"1.0.0","tractionLabel":"1.4K downloads"},"release":{"evidence":{"source":"CLAWHUB","verified":false,"confidence":"medium","updatedAt":"2026-10-10T15:42:59.533Z","emptyReason":null},"lastUpdatedAt":"2026-10-10T15:42:59.534Z","lastCrawledAt":"2026-10-10T15:42:59.533Z","lastIndexedAt":null,"nextCrawlAt":"2026-10-11T15:42:59.533Z","lastVerifiedAt":null,"highlights":[{"version":"1.0.0","createdAt":"2026-04-21T08:35:01.688Z","changelog":"首次发布：基于 5 篇论文 3 轮迭代的 SOP v2 经验，包含翻译流程、校验脚本、平台兼容性踩坑经验","fileCount":6,"zipByteSize":10030}]},"execution":{"evidence":{"source":"CLAWHUB","verified":false,"confidence":"low","updatedAt":null,"emptyReason":"No published capability contract is available yet."},"installCommand":"clawhub skill install s17czwytgebtz5mxnn62cn9zth855nvb:paper-translation","setupComplexity":"low","setupSteps":["Setup complexity is LOW. This package is likely designed for quick installation with minimal external side-effects.","Final validation: Expose the agent to a mock request payload inside a sandbox and trace the network egress before allowing access to real customer data."],"contract":{"contractStatus":"missing","authModes":[],"requires":[],"forbidden":[],"supportsMcp":false,"supportsA2a":false,"supportsStreaming":false,"inputSchemaRef":null,"outputSchemaRef":null,"dataRegion":null,"contractUpdatedAt":null,"sourceUpdatedAt":null,"freshnessSeconds":null},"invocationGuide":{"preferredApi":{"snapshotUrl":"https://www.xpersona.co/api/v1/agents/clawhub-fandywang87-paper-translation/snapshot","contractUrl":"https://www.xpersona.co/api/v1/agents/clawhub-fandywang87-paper-translation/contract","trustUrl":"https://www.xpersona.co/api/v1/agents/clawhub-fandywang87-paper-translation/trust"},"curlExamples":["curl -s \"https://www.xpersona.co/api/v1/agents/clawhub-fandywang87-paper-translation/snapshot\"","curl -s \"https://www.xpersona.co/api/v1/agents/clawhub-fandywang87-paper-translation/contract\"","curl -s \"https://www.xpersona.co/api/v1/agents/clawhub-fandywang87-paper-translation/trust\""],"jsonRequestTemplate":{"query":"summarize this repo","constraints":{"maxLatencyMs":2000,"protocolPreference":["OPENCLEW"]}},"jsonResponseTemplate":{"ok":true,"result":{"summary":"...","confidence":0.9},"meta":{"source":"CLAWHUB","generatedAt":"2026-10-10T21:53:42.675Z"}},"retryPolicy":{"maxAttempts":3,"backoffMs":[500,1500,3500],"retryableConditions":["HTTP_429","HTTP_503","NETWORK_TIMEOUT"]}},"endpoints":{"dossierUrl":"https://www.xpersona.co/api/v1/agents/clawhub-fandywang87-paper-translation/dossier","snapshotUrl":"https://www.xpersona.co/api/v1/agents/clawhub-fandywang87-paper-translation/snapshot","contractUrl":"https://www.xpersona.co/api/v1/agents/clawhub-fandywang87-paper-translation/contract","trustUrl":"https://www.xpersona.co/api/v1/agents/clawhub-fandywang87-paper-translation/trust"}},"reliability":{"evidence":{"source":"runtime-metrics","verified":false,"confidence":"low","updatedAt":null,"emptyReason":"No trust, reliability, or runtime telemetry is available."},"trust":{"status":"unavailable","handshakeStatus":"UNKNOWN","verificationFreshnessHours":null,"reputationScore":null,"p95LatencyMs":null,"successRate30d":null,"fallbackRate":null,"attempts30d":null,"trustUpdatedAt":null,"trustConfidence":"unknown","sourceUpdatedAt":null,"freshnessSeconds":null},"decisionGuardrails":{"doNotUseIf":["Contract metadata is missing or unavailable for deterministic execution."],"safeUseWhen":[],"riskFlags":["missing_or_unavailable_contract","trust_data_unavailable","schema_references_missing"],"operationalConfidence":"low"},"executionMetrics":{"observedLatencyMsP50":null,"observedLatencyMsP95":null,"estimatedCostUsd":null,"uptime30d":null,"rateLimitRpm":null,"rateLimitBurst":null,"lastVerifiedAt":null,"verificationSource":null},"runtimeMetrics":{"successRate":null,"avgLatencyMs":null,"avgCostUsd":null,"hallucinationRate":null,"retryRate":null,"disputeRate":null,"p50Latency":null,"p95Latency":null,"lastUpdated":null}},"benchmarks":{"evidence":{"source":"no-benchmark-data","verified":false,"confidence":"low","updatedAt":null,"emptyReason":"No benchmark suites or observed failure patterns are available."},"suites":[],"failurePatterns":[]},"artifacts":{"evidence":{"source":"CLAWHUB","verified":false,"confidence":"high","updatedAt":"2026-10-10T15:42:59.534Z","emptyReason":null},"readme":"Skill: 论文精读翻译\n\nOwner: fandywang87\n\nSummary: ArXiv 论文精读级中文翻译，同步到 IMA 知识库 + 腾讯文档。 基于 5 篇论文（MDL/Kunlun/OneTrans/RankMixer/MixFormer）3 轮迭代实战经验。 触发场景：翻译论文、翻译 arxiv、论文精读、论文中文翻译、paper translation、 translate p...\n\nTags: latest:1.0.0\n\nVersion history:\n\nv1.0.0 | 2026-04-21T08:35:01.688Z | user\n\n首次发布：基于 5 篇论文 3 轮迭代的 SOP v2 经验，包含翻译流程、校验脚本、平台兼容性踩坑经验\n\nArchive index:\n\nArchive v1.0.0: 6 files, 10030 bytes\n\nFiles: references/iteration-history.md (1227b), references/platform-compat.md (3350b), scripts/validate_translation.py (5553b), skill-card.md (2255b), SKILL.md (5194b), _meta.json (136b)\n\nFile v1.0.0:SKILL.md\n\n---\nname: paper-translation\ndescription: >\n  ArXiv 论文精读级中文翻译，同步到 IMA 知识库 + 腾讯文档。\n  基于 5 篇论文（MDL/Kunlun/OneTrans/RankMixer/MixFormer）3 轮迭代实战经验。\n  触发场景：翻译论文、翻译 arxiv、论文精读、论文中文翻译、paper translation、\n  translate paper、翻译这篇论文、帮我翻译、精读翻译、论文全文翻译、\n  把这篇论文翻译成中文、翻译 arxiv 论文到知识库。\n---\n\n# ArXiv 论文精读翻译\n\nBase directory for this skill: `{SKILL_DIR}`\n\n将 ArXiv 论文逐段翻译成中文，生成双版本 Markdown（IMA + 腾讯文档），并上传到两个平台。\n\n## 三条铁律\n\n1. **完整翻译不精简** — 逐段翻译每个 paragraph，不遗漏任何论证细节。大模型倾向于\"帮你归纳\"，但用户要的是精读级翻译。\n2. **译注显式标记** — 大模型解读必须用 `> **[译注]**：...` 引用块，绝不混入原文翻译。\n3. **简称首次标全称，后续直接用** — 首次出现标注全称并核对原文，后续不再展开。避免错误展开（如 TA=Target Attention 被误写为 Transformer Aggregator）。\n\n## 标准 6 步流程\n\n### Step 1: 获取原文\n\n```\nweb_fetch https://arxiv.org/html/<id>v<n>\n```\n\n- 只 fetch **一次**，节省 token\n- 同步下载图片：`curl -sL -o x{n}.png https://arxiv.org/html/<paper_id>/x{n}.png`\n- 下载后检查文件大小，相同大小的异常文件（404 垃圾响应）删除\n\n### Step 2: 翻译生成\n\n- **逐段翻译**，不做精简\n- 首行元信息：原标题、arxiv 链接、年月、机构、翻译辅助大模型名称\n- 简称首次出现标全称（核对原文），后续用简称\n- 译注用 `> **[译注]**：...` 格式\n- 结构化排版：多级标题 + 列表 + 加粗 + 表格 + 引用块\n- 公式保留 LaTeX；**`\\bm` 全部替换为 `\\boldsymbol`**\n- 图表按原文顺序插入**所在章节标题之后、小节正文之前**\n- 参考文献完整列出\n\n**表格处理策略**：\n- 简单表格 → Markdown 表格重写（可搜索/编辑）\n- 复杂表格（合并单元格/特殊排版）→ PyMuPDF 从 PDF 截取\n\n### Step 3: 自动化校验\n\n翻译完成后，运行校验脚本：\n\n```bash\npython3 {SKILL_DIR}/scripts/validate_translation.py <markdown_file>\n```\n\n校验项：\n\n| 检查项 | 标准 |\n|--------|------|\n| 章节完整性 | 包含：摘要/引言/相关工作/方法/实验/结论/参考文献 |\n| LaTeX 兼容性 | `\\bm` 出现次数 = 0 |\n| 译注标记 | 数量 > 0，格式为 `> **[译注]**` |\n| 参考文献 | 条数列出供人工核对 |\n| 图片链接 | 外链格式正确 |\n\n### Step 4: 生成两版 Markdown\n\n- **IMA 版**：图片用 arxiv 外链 URL / base64 data URI\n- **腾讯文档版**：用脚本从 IMA 版自动替换图片链接为 image_id\n\n图片上传流程：\n1. `curl -sL -o x{n}.png https://arxiv.org/html/<paper_id>/x{n}.png` 下载\n2. 腾讯文档：`mcporter call tencent-docs upload_image` → 拿 image_id\n3. IMA：直接用 arxiv 外链 URL\n\n详见 [references/platform-compat.md](references/platform-compat.md)\n\n### Step 5: 上传 IMA 知识库\n\n```bash\n# create_media → COS 上传 → add_knowledge（media_type=7 = Markdown）\n# 如遇 code=220030（限流），sleep 15s 重试，cos_key 仍有效\n```\n\n### Step 6: 上传腾讯文档\n\n```bash\nTITLE=\"【YYYY.MM｜组织】XXX 中文翻译\"  # 必须 ≤36 字符\njq -n --arg title \"$TITLE\" --rawfile mdx \"$FILE\" --arg cf \"markdown\" \\\n  '{title:$title, mdx:$mdx, content_format:$cf}' > /tmp/args.json\nmcporter call tencent-docs create_smartcanvas_by_mdx --args \"$(cat /tmp/args.json)\"\n```\n\nmcporter 传大参数**不支持 `--args-file`**，必须用 `--args \"$(cat file.json)\"`。\n\n## 命名规范（强制）\n\n| 平台 | 格式 | 约束 |\n|------|------|------|\n| 腾讯文档标题 | `【YYYY.MM｜组织】XXX 中文翻译` | ≤36 字符（按字符数，非字节） |\n| IMA 文件名 | `【YYYY.MM｜组织】XXX 中文翻译.md` | 同名加 .md |\n\n- 两平台**必须完全一致**，不加 `v2`/`图文版` 等后缀\n- 示例：`【2026.02｜ByteDance】MixFormer 中文翻译`\n\n## 翻译后 Checklist\n\n完成翻译后逐项确认：\n\n- [ ] `grep -c '\\\\bm'` = 0\n- [ ] 简称首次出现已标全称且正确\n- [ ] 译注均用 `> **[译注]**：...` 格式\n- [ ] 图表位置与原文章节顺序一致\n- [ ] 参考文献条数与原文一致\n- [ ] IMA 版和腾讯文档版图片格式各自正确\n- [ ] 两平台文件名/标题完全一致\n- [ ] 首行包含论文元信息（标题/链接/年月/机构/大模型名称）\n\n## 效率优化\n\n- 只 `web_fetch` 一次原文（节省 token）\n- 直接生成最终版，不生成中间草稿（减少 50%+ 工具调用）\n- 图片下载 + 上传并行执行\n- 用脚本自动从 IMA 版生成腾讯文档版\n- 自动化校验脚本在上传前拦截格式问题\n\n## 参考文档\n\n- **踩坑经验 + 平台兼容性**：[references/platform-compat.md](references/platform-compat.md) — LaTeX 兼容、图片跨平台、API 限流等详细说明\n- **迭代历史**：[references/iteration-history.md](references/iteration-history.md) — MixFormer v1→v3 的完整教训记录\n\nFile v1.0.0:_meta.json\n\n{\n  \"ownerId\": \"kn7fgq1zdyxj5g08emhtwsdzc1855zqb\",\n  \"slug\": \"paper-translation\",\n  \"version\": \"1.0.0\",\n  \"publishedAt\": 1776760501688\n}\n\nFile v1.0.0:references/iteration-history.md\n\n# 翻译迭代历史（MixFormer 案例）\n\n## 版本演进\n\n| 版本 | 主要问题 | 改进措施 |\n|------|---------|---------|\n| **v1** | 过度精简、译注未标记、表格用截图不可搜索、图片位置随意 | 用户反馈 5 大改进点 |\n| **v2** | 逐段翻译、译注标记、表格改 Markdown、自动化校验 | 图片位置仍有偏差、译注覆盖不够 |\n| **v3** | 精确核对图片章节位置、19 处译注（vs v2 的 9 处）、TA 全称修正 | 趋于成熟，形成 SOP v2 |\n\n## 关键教训\n\n1. **翻译质量需要 2-3 轮迭代**才能达到用户满意\n2. **v1 的根因**：大模型自作主张精简 + 译注混入正文 + 简称展开错误\n3. **v2 的不足**：大方向正确但细节不够（图片位置偏差、译注覆盖不充分）\n4. **v3 的突破**：在细节上达标，形成可复用的 SOP\n\n## 已翻译论文清单\n\n| 论文 | 机构 | 年月 | 特殊注意 |\n|------|------|------|---------|\n| MDL | ByteDance | 2026.02 | — |\n| Kunlun | Meta（非百度） | 2026.02 | 机构归属易搞错 |\n| OneTrans | ByteDance | 2025.10 | 图片 x5-x8 为 404 |\n| RankMixer | ByteDance | 2025.07 | — |\n| MixFormer | ByteDance | 2026.02 | 经历 v1→v3 三轮迭代 |\n\nFile v1.0.0:references/platform-compat.md\n\n# 平台兼容性与踩坑经验\n\n## LaTeX 公式兼容性\n\n| 问题 | 解决方案 |\n|------|---------|\n| 腾讯文档不支持 `\\bm{...}`（bm 宏包） | 全部替换为 `\\boldsymbol{...}` |\n| `\\bold` / `\\mathbold` 不兼容 | 统一到 `\\boldsymbol` |\n| 罕见宏包命令 | 避免使用，仅 `\\mathbb{R}` 等标准宏安全 |\n\n**强制流程**：翻译完成后执行 `grep -c '\\\\bm' file.md`，结果必须为 0。\n\n## 图片跨平台兼容性\n\n| 平台 | 支持 | 不支持 |\n|------|------|--------|\n| IMA 知识库 | ArXiv 外链 URL、base64 data URI | 本地相对路径（上传后失效） |\n| 腾讯文档 | `upload_image` 返回的 image_id | HTTP/HTTPS 外链 |\n\n**标准图片上传流程**：\n\n```bash\n# 1. 下载\ncurl -sL -o x{n}.png https://arxiv.org/html/<paper_id>/x{n}.png\n\n# 2. 腾讯文档：上传拿 image_id\nmcporter call tencent-docs upload_image --args '{\"file_name\":\"x1.png\",\"image_base64\":\"<base64>\"}'\n\n# 3. IMA：直接用外链\n![Fig](https://arxiv.org/html/<paper_id>/x1.png)\n```\n\n**图片 404 排查**：下载后检查文件大小，多个文件大小完全相同通常为 404 垃圾响应，应删除。\n\n## 图表插入位置\n\n图/表插入在**所在章节标题之后、小节正文之前**（与原文顺序一致）。\n\n- Figure 2 应放在 §3.4 章节入口，而非 §3.4.1 之后\n- Table 1/2 放在章节标题后即可\n- 避免图表打断论证链条\n\n## 腾讯文档 API 踩坑\n\n### 标题长度限制\n\n`create_smartcanvas_by_mdx` 的 `title` 字段**限制 36 字符**（按字符数计算，非字节）。\n\n超长报错：`business 400001: title length exceeds 36 characters`\n\n### mcporter 传大参数\n\n`mcporter call` **不支持 `--args-file`**。可靠写法：\n\n```bash\n# 先生成 JSON 文件\njq -n --arg title \"$TITLE\" --rawfile mdx \"$FILE\" --arg cf \"markdown\" \\\n  '{title:$title, mdx:$mdx, content_format:$cf}' > /tmp/args.json\n\n# 再用 cat 传入\nmcporter call tencent-docs create_smartcanvas_by_mdx --args \"$(cat /tmp/args.json)\"\n```\n\n### 授权流程\n\n`setup.sh` 在 CodeBuddy 独立 shell 环境下后台进程会被回收。解决方案：绕开 setup.sh，用同步轮询获取 token：\n\n1. `openssl rand -hex 8` 生成 code\n2. 展示授权链接给用户\n3. `curl` 同步轮询 token（每 10 秒一次，最多 3 分钟）\n4. `mcporter config add` 注册 token\n\n## IMA 知识库 API 踩坑\n\n### add_knowledge 限流\n\n短时间内多次调用返回 `code=220030`（表面\"没有权限\"，实际是限流）。\n\n**解决方案**：sleep 15 秒后重试，已上传的 cos_key 仍有效，无需重新 create_media。\n\n### 图片不支持直传\n\n`create_media` 的 `media_type=9`（图片）返回 220030。图片只能嵌入 Markdown（base64 data URI 或外链 URL）。\n\n## 翻译内容质量常见问题\n\n| 问题 | 表现 | 预防 |\n|------|------|------|\n| 简称全称搞错 | TA=Target Attention 错写为 Transformer Aggregator | 首次出现标全称并核对原文 |\n| 默认精简原文 | 完整论证压缩成摘要式描述 | 逐段翻译，不做默认精简 |\n| 模型解读混入正文 | 评论和翻译混在一起 | 强制 `> **[译注]**：...` 引用块 |\n| 事实性错误 | Kunlun 归属错写为百度（实际 Meta） | 元信息必须从原文/搜索确认 |\n| PDF 截图文件名搞反 | table1.png 实际是 Table 2 | 保存后按内容核对文件名 |\n\nFile v1.0.0:skill-card.md\n\n## Description:\n\nArXiv 论文精读级中文翻译，同步到 IMA 知识库和腾讯文档，并提供翻译校验流程。\n\nThis skill is ready for commercial/non-commercial use.\n\n## Publisher:\n\n[fandywang87](https://clawhub.ai/user/fandywang87)\n\n### License/Terms of Use:\n\nMIT-0\n\n## Use Case:\n\nExternal users and developers use this skill to produce complete Chinese translations of public ArXiv papers, preserve paper structure and formulas, add clearly marked translator notes, validate the Markdown, and prepare IMA and Tencent Docs versions.\n\n### Deployment Geography for Use:\n\nGlobal\n\n## Known Risks and Mitigations:\n\nRisk: Persistent uploads to Tencent Docs and IMA can expose translated content outside the local agent session.\n\nMitigation: Add an explicit user confirmation step before upload and use the workflow only for public ArXiv papers or documents approved for those platforms.\n\nRisk: The workflow uses a broad trigger surface for paper translation tasks.\n\nMitigation: Confirm the target paper, destination platforms, and upload intent before fetching, translating, or publishing content.\n\nRisk: The temporary file pattern `/tmp/args.json` can collide with other processes or leave upload payloads on disk.\n\nMitigation: Use a unique restricted temporary file for each run and delete it immediately after the platform call completes.\n\n## Reference(s):\n\n- [平台兼容性与踩坑经验](references/platform-compat.md)\n- [翻译迭代历史（MixFormer 案例）](references/iteration-history.md)\n- [ClawHub skill page](https://clawhub.ai/fandywang87/skills/paper-translation)\n\n## Skill Output:\n\n**Output Type(s):** [Markdown, Shell commands, Configuration, Guidance]\n\n**Output Format:** [Markdown with inline shell commands and validation guidance]\n\n**Output Parameters:** [1D]\n\n**Other Properties Related to Output:** [Produces IMA and Tencent Docs oriented Markdown versions and a validation checklist.]\n\n## Skill Version(s):\n\n1.0.0 (source: server release evidence)\n\n## Ethical Considerations:\n\nUsers should evaluate whether this skill is appropriate for their environment, review any generated or modified files before relying on them, and apply their organization's safety, security, and compliance requirements before deployment.","readmeExcerpt":"Skill: 论文精读翻译 Owner: fandywang87 Summary: ArXiv 论文精读级中文翻译，同步到 IMA 知识库 + 腾讯文档。 基于 5 篇论文（MDL/Kunlun/OneTrans/RankMixer/MixFormer）3 轮迭代实战经验。 触发场景：翻译论文、翻译 arxiv、论文精读、论文中文翻译、paper translation、 translate p... Tags: latest:1.0.0 Version history: v1.0.0 | 2026-04-21T08:35:01.688Z | user 首次发布：基于 5 篇论文 3 轮迭代的 SOP v2 经验，包含翻译流程、校验脚本、平台兼容性踩坑经验 Archive index: Archive v1.0.0: 6 files, 10030 bytes Files: references/iteration-history","codeSnippets":[],"executableExamples":[{"language":"text","snippet":"web_fetch https://arxiv.org/html/<id>v<n>"},{"language":"bash","snippet":"python3 {SKILL_DIR}/scripts/validate_translation.py <markdown_file>"},{"language":"bash","snippet":"# create_media → COS 上传 → add_knowledge（media_type=7 = Markdown）\n# 如遇 code=220030（限流），sleep 15s 重试，cos_key 仍有效"},{"language":"bash","snippet":"TITLE=\"【YYYY.MM｜组织】XXX 中文翻译\"  # 必须 ≤36 字符\njq -n --arg title \"$TITLE\" --rawfile mdx \"$FILE\" --arg cf \"markdown\" \\\n  '{title:$title, mdx:$mdx, content_format:$cf}' > /tmp/args.json\nmcporter call tencent-docs create_smartcanvas_by_mdx --args \"$(cat /tmp/args.json)\""},{"language":"bash","snippet":"curl -sL -o x{n}.png https://arxiv.org/html/<paper_id>/x{n}.png"},{"language":"bash","snippet":"# 1. 下载\ncurl -sL -o x{n}.png https://arxiv.org/html/<paper_id>/x{n}.png\n\n# 2. 腾讯文档：上传拿 image_id\nmcporter call tencent-docs upload_image --args '{\"file_name\":\"x1.png\",\"image_base64\":\"<base64>\"}'\n\n# 3. IMA：直接用外链\n![Fig](https://arxiv.org/html/<paper_id>/x1.png)"}],"parameters":null,"dependencies":[],"permissions":[],"extractedFiles":[{"path":"SKILL.md","content":"---\nname: paper-translation\ndescription: >\n  ArXiv 论文精读级中文翻译，同步到 IMA 知识库 + 腾讯文档。\n  基于 5 篇论文（MDL/Kunlun/OneTrans/RankMixer/MixFormer）3 轮迭代实战经验。\n  触发场景：翻译论文、翻译 arxiv、论文精读、论文中文翻译、paper translation、\n  translate paper、翻译这篇论文、帮我翻译、精读翻译、论文全文翻译、\n  把这篇论文翻译成中文、翻译 arxiv 论文到知识库。\n---\n\n# ArXiv 论文精读翻译\n\nBase directory for this skill: `{SKILL_DIR}`\n\n将 ArXiv 论文逐段翻译成中文，生成双版本 Markdown（IMA + 腾讯文档），并上传到两个平台。\n\n## 三条铁律\n\n1. **完整翻译不精简** — 逐段翻译每个 paragraph，不遗漏任何论证细节。大模型倾向于\"帮你归纳\"，但用户要的是精读级翻译。\n2. **译注显式标记** — 大模型解读必须用 `> **[译注]**：...` 引用块，绝不混入原文翻译。\n3. **简称首次标全称，后续直接用** — 首次出现标注全称并核对原文，后续不再展开。避免错误展开（如 TA=Target Attention 被误写为 Transformer Aggregator）。\n\n## 标准 6 步流程\n\n### Step 1: 获取原文\n\n```\nweb_fetch https://arxiv.org/html/<id>v<n>\n```\n\n- 只 fetch **一次**，节省 token\n- 同步下载图片：`curl -sL -o x{n}.png https://arxiv.org/html/<paper_id>/x{n}.png`\n- 下载后检查文件大小，相同大小的异常文件（404 垃圾响应）删除\n\n### Step 2: 翻译生成\n\n- **逐段翻译**，不做精简\n- 首行元信息：原标题、arxiv 链接、年月、机构、翻译辅助大模型名称\n- 简称首次出现标全称（核对原文），后续用简称\n- 译注用 `> **[译注]**：...` 格式\n- 结构化排版：多级标题 + 列表 + 加粗 + 表格 + 引用块\n- 公式保留 LaTeX；**`\\bm` 全部替换为 `\\boldsymbol`**\n- 图表按原文顺序插入**所在章节标题之后、小节正文之前**\n- 参考文献完整列出\n\n**表格处理策略**：\n- 简单表格 → Markdown 表格重写（可搜索/编辑）\n- 复杂表格（合并单元格/特殊排版）→ PyMuPDF 从 PDF 截取\n\n### Step 3: 自动化校验\n\n翻译完成后，运行校验脚本：\n\n```bash\npython3 {SKILL_DIR}/scripts/validate_translation.py <markdown_file>\n```\n\n校验项：\n\n| 检查项 | 标准 |\n|--------|------|\n| 章节完整性 | 包含：摘要/引言/相关工作/方法/实验/结论/参考文献 |\n| LaTeX 兼容性 | `\\bm` 出现次数 = 0 |\n| 译注标记 | 数量 > 0，格式为 `> **[译注]**` |\n| 参考文献 | 条数列出供人工核对 |\n| 图片链接 | 外链格式正确 |\n\n### Step 4: 生成两版 Markdown\n\n- **IMA 版**：图片用 arxiv 外链 URL / base64 data URI\n- **腾讯文档版**：用脚本从 IMA 版自动替换图片链接为 image_id\n\n图片上传流程：\n1. `curl -sL -o x{n}.png https://arxiv.org/html/<paper_id>/x{n}.png` 下载\n2. 腾讯文档：`mcporter call tencent-docs upload_image` → 拿 image_id\n3. IMA：直接用 arxiv 外链 URL\n\n详见 [references/platform-compat.md](references/platform-compat.md)\n\n### Step 5: 上传 IMA 知识库\n\n```bash\n# create_media → COS 上传 → add_knowledge（media_type=7 = Markdown）\n# 如遇 code=220030（限流），sleep 15s 重试，cos_key 仍有效\n```\n\n### Step 6: 上传腾讯文档\n\n```bash\nTITLE=\"【YYYY.MM｜组织】XXX 中文翻译\"  # 必须 ≤36 字符\njq -n --arg title \"$TITLE\" --rawfile mdx \"$FILE\" --arg cf \"markdown\" \\\n  '{title:$title, mdx:$mdx, content_format:$cf}' > /tmp/args.json\nmcporter call tencent-docs create_smartcanvas_by_mdx --args \"$(cat /tmp/args.json)\"\n```\n\nmcporter 传大参数**不支持 `--args-file`**，必须用 `--args \"$(cat file.json)\"`。\n\n## 命名规范（强制）\n\n| 平台 | 格式 | 约束 |\n|------|------|------|\n| 腾讯文档标题 | `【YYYY.MM｜组织】XXX 中文翻译` | ≤36 字符（按字符数，非字节） |\n| IMA 文件名 | `【YYYY.MM｜组织】XXX 中文翻译.md` | 同名加 .md |\n\n- 两平台**必须完全一致**，不加 `v2`/`图文版` 等后缀\n- 示例：`【2026.02｜ByteDance】MixFormer 中文翻译`\n\n## 翻译后 Checklist\n\n完成翻译后逐项确认：\n\n- [ ] `grep -c '\\\\bm'` = 0\n- [ ] 简称首次出现已标全称且正确\n- [ ] 译注均用 `> **[译注]**：...` 格式\n- [ ] 图表位置与原文章节顺序一致\n- [ ] 参考文献条数与原文一致\n- [ ] IMA 版和腾讯文档版图片格式各自正确\n- [ ] 两平台文件名/标题完全一致\n- [ ] 首行包含论文元信息（标题/链接/年月/机构/大模型名称）\n\n## 效率优化\n\n- 只 `web_fetch` 一次原文（节省 token）\n- 直接生成最终版，不生成中间草稿（减少 50%+ 工具调用）\n- 图片下载 + 上传并行执行\n- 用脚本自动从 IMA 版生成腾讯文档版\n- 自动化校验脚本在上传前拦截格式问题\n\n## 参考文档\n\n- **踩坑经验 + 平台兼容性**：[references/platform-compat.md](references/pla"},{"path":"_meta.json","content":"{\n  \"ownerId\": \"kn7fgq1zdyxj5g08emhtwsdzc1855zqb\",\n  \"slug\": \"paper-translation\",\n  \"version\": \"1.0.0\",\n  \"publishedAt\": 1776760501688\n}"},{"path":"references/iteration-history.md","content":"# 翻译迭代历史（MixFormer 案例）\n\n## 版本演进\n\n| 版本 | 主要问题 | 改进措施 |\n|------|---------|---------|\n| **v1** | 过度精简、译注未标记、表格用截图不可搜索、图片位置随意 | 用户反馈 5 大改进点 |\n| **v2** | 逐段翻译、译注标记、表格改 Markdown、自动化校验 | 图片位置仍有偏差、译注覆盖不够 |\n| **v3** | 精确核对图片章节位置、19 处译注（vs v2 的 9 处）、TA 全称修正 | 趋于成熟，形成 SOP v2 |\n\n## 关键教训\n\n1. **翻译质量需要 2-3 轮迭代**才能达到用户满意\n2. **v1 的根因**：大模型自作主张精简 + 译注混入正文 + 简称展开错误\n3. **v2 的不足**：大方向正确但细节不够（图片位置偏差、译注覆盖不充分）\n4. **v3 的突破**：在细节上达标，形成可复用的 SOP\n\n## 已翻译论文清单\n\n| 论文 | 机构 | 年月 | 特殊注意 |\n|------|------|------|---------|\n| MDL | ByteDance | 2026.02 | — |\n| Kunlun | Meta（非百度） | 2026.02 | 机构归属易搞错 |\n| OneTrans | ByteDance | 2025.10 | 图片 x5-x8 为 404 |\n| RankMixer | ByteDance | 2025.07 | — |\n| MixFormer | ByteDance | 2026.02 | 经历 v1→v3 三轮迭代 |"},{"path":"references/platform-compat.md","content":"# 平台兼容性与踩坑经验\n\n## LaTeX 公式兼容性\n\n| 问题 | 解决方案 |\n|------|---------|\n| 腾讯文档不支持 `\\bm{...}`（bm 宏包） | 全部替换为 `\\boldsymbol{...}` |\n| `\\bold` / `\\mathbold` 不兼容 | 统一到 `\\boldsymbol` |\n| 罕见宏包命令 | 避免使用，仅 `\\mathbb{R}` 等标准宏安全 |\n\n**强制流程**：翻译完成后执行 `grep -c '\\\\bm' file.md`，结果必须为 0。\n\n## 图片跨平台兼容性\n\n| 平台 | 支持 | 不支持 |\n|------|------|--------|\n| IMA 知识库 | ArXiv 外链 URL、base64 data URI | 本地相对路径（上传后失效） |\n| 腾讯文档 | `upload_image` 返回的 image_id | HTTP/HTTPS 外链 |\n\n**标准图片上传流程**：\n\n```bash\n# 1. 下载\ncurl -sL -o x{n}.png https://arxiv.org/html/<paper_id>/x{n}.png\n\n# 2. 腾讯文档：上传拿 image_id\nmcporter call tencent-docs upload_image --args '{\"file_name\":\"x1.png\",\"image_base64\":\"<base64>\"}'\n\n# 3. IMA：直接用外链\n![Fig](https://arxiv.org/html/<paper_id>/x1.png)\n```\n\n**图片 404 排查**：下载后检查文件大小，多个文件大小完全相同通常为 404 垃圾响应，应删除。\n\n## 图表插入位置\n\n图/表插入在**所在章节标题之后、小节正文之前**（与原文顺序一致）。\n\n- Figure 2 应放在 §3.4 章节入口，而非 §3.4.1 之后\n- Table 1/2 放在章节标题后即可\n- 避免图表打断论证链条\n\n## 腾讯文档 API 踩坑\n\n### 标题长度限制\n\n`create_smartcanvas_by_mdx` 的 `title` 字段**限制 36 字符**（按字符数计算，非字节）。\n\n超长报错：`business 400001: title length exceeds 36 characters`\n\n### mcporter 传大参数\n\n`mcporter call` **不支持 `--args-file`**。可靠写法：\n\n```bash\n# 先生成 JSON 文件\njq -n --arg title \"$TITLE\" --rawfile mdx \"$FILE\" --arg cf \"markdown\" \\\n  '{title:$title, mdx:$mdx, content_format:$cf}' > /tmp/args.json\n\n# 再用 cat 传入\nmcporter call tencent-docs create_smartcanvas_by_mdx --args \"$(cat /tmp/args.json)\"\n```\n\n### 授权流程\n\n`setup.sh` 在 CodeBuddy 独立 shell 环境下后台进程会被回收。解决方案：绕开 setup.sh，用同步轮询获取 token：\n\n1. `openssl rand -hex 8` 生成 code\n2. 展示授权链接给用户\n3. `curl` 同步轮询 token（每 10 秒一次，最多 3 分钟）\n4. `mcporter config add` 注册 token\n\n## IMA 知识库 API 踩坑\n\n### add_knowledge 限流\n\n短时间内多次调用返回 `code=220030`（表面\"没有权限\"，实际是限流）。\n\n**解决方案**：sleep 15 秒后重试，已上传的 cos_key 仍有效，无需重新 create_media。\n\n### 图片不支持直传\n\n`create_media` 的 `media_type=9`（图片）返回 220030。图片只能嵌入 Markdown（base64 data URI 或外链 URL）。\n\n## 翻译内容质量常见问题\n\n| 问题 | 表现 | 预防 |\n|------|------|------|\n| 简称全称搞错 | TA=Target Attention 错写为 Transformer Aggregator | 首次出现标全称并核对原文 |\n| 默认精简原文 | 完整论证压缩成摘要式描述 | 逐段翻译，不做默认精简 |\n| 模型解读混入正文 | 评论和翻译混在一起 | 强制 `> **[译注]**：...` 引用块 |\n| 事实性错误 | Kunlun 归属错写为百度（实际 Meta） | 元信息必须从原文/搜索确认 |\n| PDF 截图文件名搞反 | table1.png 实际是 Table 2 | 保存后按内容核对文件名 |"},{"path":"skill-card.md","content":"## Description:\n\nArXiv 论文精读级中文翻译，同步到 IMA 知识库和腾讯文档，并提供翻译校验流程。\n\nThis skill is ready for commercial/non-commercial use.\n\n## Publisher:\n\n[fandywang87](https://clawhub.ai/user/fandywang87)\n\n### License/Terms of Use:\n\nMIT-0\n\n## Use Case:\n\nExternal users and developers use this skill to produce complete Chinese translations of public ArXiv papers, preserve paper structure and formulas, add clearly marked translator notes, validate the Markdown, and prepare IMA and Tencent Docs versions.\n\n### Deployment Geography for Use:\n\nGlobal\n\n## Known Risks and Mitigations:\n\nRisk: Persistent uploads to Tencent Docs and IMA can expose translated content outside the local agent session.\n\nMitigation: Add an explicit user confirmation step before upload and use the workflow only for public ArXiv papers or documents approved for those platforms.\n\nRisk: The workflow uses a broad trigger surface for paper translation tasks.\n\nMitigation: Confirm the target paper, destination platforms, and upload intent before fetching, translating, or publishing content.\n\nRisk: The temporary file pattern `/tmp/args.json` can collide with other processes or leave upload payloads on disk.\n\nMitigation: Use a unique restricted temporary file for each run and delete it immediately after the platform call completes.\n\n## Reference(s):\n\n- [平台兼容性与踩坑经验](references/platform-compat.md)\n- [翻译迭代历史（MixFormer 案例）](references/iteration-history.md)\n- [ClawHub skill page](https://clawhub.ai/fandywang87/skills/paper-translation)\n\n## Skill Output:\n\n**Output Type(s):** [Markdown, Shell commands, Configuration, Guidance]\n\n**Output Format:** [Markdown with inline shell commands and validation guidance]\n\n**Output Parameters:** [1D]\n\n**Other Properties Related to Output:** [Produces IMA and Tencent Docs oriented Markdown versions and a validation checklist.]\n\n## Skill Version(s):\n\n1.0.0 (source: server release evidence)\n\n## Ethical Considerations:\n\nUsers should evaluate whether this skill is appropriate for their environment, review any generated or modified files before relying on them, and apply their organization's safety, security, and compliance requirements before deployment."}],"languages":[],"docsSourceLabel":"CLAWHUB","editorialOverview":"ArXiv 论文精读级中文翻译，同步到 IMA 知识库 + 腾讯文档。 基于 5 篇论文（MDL/Kunlun/OneTrans/RankMixer/MixFormer）3 轮迭代实战经验。 触发场景：翻译论文、翻译 arxiv、论文精读、论文中文翻译、paper translation、 translate p... Skill: 论文精读翻译 Owner: fandywang87 Summary: ArXiv 论文精读级中文翻译，同步到 IMA 知识库 + 腾讯文档。 基于 5 篇论文（MDL/Kunlun/OneTrans/RankMixer/MixFormer）3 轮迭代实战经验。 触发场景：翻译论文、翻译 arxiv、论文精读、论文中文翻译、paper translation、 translate p... Tags: latest:1.0.0 Version history: v1.0.0 | 2026-04-21T08:35:01.688Z | user 首次发布：基于 5 篇论文 3 轮迭代的 SOP v2 经验，包含翻译流程、校验脚本、平台兼容性踩坑经验 Archive index: Archive v1.0.0: 6 files, 10030 bytes Files: references/iteration-history","editorialQuality":{"score":100,"threshold":65,"status":"ready","wordCount":708,"uniquenessScore":62,"reasons":[]}},"media":{"evidence":{"source":"no-media","verified":false,"confidence":"low","updatedAt":"2026-10-10T15:42:59.534Z","emptyReason":"No screenshots, media assets, or demo links are available."},"primaryImageUrl":null,"mediaAssetCount":0,"assets":[],"demoUrl":null},"ownerResources":{"evidence":{"source":"unclaimed","verified":false,"confidence":"low","updatedAt":"2026-10-10T15:42:59.534Z","emptyReason":"This page has not been claimed by the agent owner."},"hasCustomPage":false,"customPageUpdatedAt":null,"customLinks":[],"structuredLinks":{"docsUrl":null,"demoUrl":null,"supportUrl":null,"pricingUrl":null,"statusUrl":null},"customPage":null},"relatedAgents":{"evidence":{"source":"protocol-neighbors","verified":false,"confidence":"medium","updatedAt":"2026-10-10T21:53:42.675Z","emptyReason":null},"items":[{"id":"8ebccd8e-3863-4187-8355-c3f14e1f9edf","entityType":"agent","canonicalPath":"/agent/iofficeai-aionui","slug":"iofficeai-aionui","name":"AionUi","description":"Free, local, open-source 24/7 Cowork app and OpenClaw for Gemini CLI, Claude Code, Codex, OpenCode, Qwen Code, Goose CLI, Auggie, and more | 🌟 Star if you like it!","url":"https://github.com/iOfficeAI/AionUi","homepage":"https://www.aionui.com","source":"GITHUB_REPOS","protocols":["MCP","OPENCLAW"],"capabilities":[],"safetyScore":100,"overallRank":70,"updatedAt":"2026-10-09T19:11:12.944Z","createdAt":"2026-02-25T03:38:16.584Z","downloads":null},{"id":"b917f68a-ebff-438e-84f8-3f4b2494c0bc","entityType":"agent","canonicalPath":"/agent/activepieces-activepieces","slug":"activepieces-activepieces","name":"activepieces","description":"AI Agents & MCPs & AI Workflow Automation • (~400 MCP servers for AI agents) • AI Automation / AI Agent with MCPs • AI Workflows & AI Agents • MCPs for AI Agents","url":"https://github.com/activepieces/activepieces","homepage":"https://www.activepieces.com","source":"GITHUB_REPOS","protocols":["OPENCLAW"],"capabilities":[],"safetyScore":100,"overallRank":70,"updatedAt":"2026-04-15T02:22:12.426Z","createdAt":"2026-02-25T03:38:12.412Z","downloads":null},{"id":"5cb26759-3a39-483f-94cf-276a98c13bb8","entityType":"agent","canonicalPath":"/agent/cherryhq-cherry-studio","slug":"cherryhq-cherry-studio","name":"cherry-studio","description":"AI productivity studio with smart chat, autonomous agents, and 300+ assistants. Unified access to frontier LLMs","url":"https://github.com/CherryHQ/cherry-studio","homepage":"https://cherry-ai.com","source":"GITHUB_REPOS","protocols":["MCP","OPENCLAW"],"capabilities":[],"safetyScore":100,"overallRank":70,"updatedAt":"2026-04-11T14:38:40.986Z","createdAt":"2026-02-25T03:38:19.379Z","downloads":null},{"id":"6f6582d0-5d76-4f0f-b81d-86520247950b","entityType":"agent","canonicalPath":"/agent/copilotkit-copilotkit","slug":"copilotkit-copilotkit","name":"CopilotKit","description":"The Frontend for Agents & Generative UI. React + Angular","url":"https://github.com/CopilotKit/CopilotKit","homepage":"https://docs.copilotkit.ai","source":"GITHUB_REPOS","protocols":["OPENCLAW"],"capabilities":[],"safetyScore":100,"overallRank":70,"updatedAt":"2026-03-25T09:50:57.846Z","createdAt":"2026-02-25T03:39:14.617Z","downloads":null}],"links":{"hub":"/agent","source":"/agent/source/clawhub","protocols":[{"label":"OpenClaw","href":"/agent/protocol/openclew"}]}}}