[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"skill:01M7BTX02FS4HP5KG3PFPNDZEW":3,"skill-same:01M7BTX02FS4HP5KG3PFPNDZEW":45},{"skillNo":4,"name":5,"nameZh":6,"description":7,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":16,"categories":17,"keywords":18,"publishedAt":19,"linkOnly":20,"markdownContent":21,"markdownBody":22,"frontMatter":23,"refName":24,"commitSha":25,"skillDirectory":26,"forksCount":27,"repositoryHtmlUrl":28,"repositoryHomepageUrl":6,"ownerHtmlUrl":29,"ownerBlogUrl":6,"ownerDescription":6,"updatedAt":30,"install":31,"resources":44},"01M7BTX02FS4HP5KG3PFPNDZEW","enterprise-artifact-search",null,"Multi-hop evidence search + structured extraction over enterprise artifact datasets (docs\u002Fchats\u002Fmeetings\u002FPRs\u002FURLs), with strict question grounding, anti-distractor filtering, evidence-based entity selection, ID validation, JSON-ready outputs, and token logging that avoids unsupported installs.","linyouxue\u002Fskill-repair-benchmark","linyouxue","https:\u002F\u002Favatars.githubusercontent.com\u002Fu\u002F216151464?v=4","01M7BJEVCXKD74KH6FG9CBVSQB","林宥学","01M7BJEVE02NP217V93BKZFC0J",0,"Apache-2.0","https:\u002F\u002Fgithub.com\u002Flinyouxue\u002Fskill-repair-benchmark\u002Fblob\u002Fmain\u002Fevaluation\u002Fresults\u002FMMG2Skill-GPT52-Gold31-20260923\u002Ftasks\u002Fenterprise-information-search\u002Frepaired_skill\u002Fenterprise-artifact-search\u002FSKILL.md",[],[],"2026-11-15 14:19:17",false,"---\nname: enterprise-artifact-search\ndescription: Multi-hop evidence search + structured extraction over enterprise artifact\n  datasets (docs\u002Fchats\u002Fmeetings\u002FPRs\u002FURLs), with strict question grounding, anti-distractor\n  filtering, evidence-based entity selection, ID validation, JSON-ready outputs, and\n  token logging that avoids unsupported installs.\n---\n\n## Steps\n1. **Ground on the question text (no heuristics):**\n   - Read `\u002Froot\u002Fquestion.txt` and treat each entry (e.g., `q1`, `q2`, …) as the *only* specification for what to retrieve.\n   - For each question, explicitly extract: target product\u002Ftopic, required entity types (IDs\u002Fnames\u002FURLs\u002Fetc.), and any constraints (“author”, “key reviewers”, “competitor demos”, “from Slack”, “latest”, etc.).\n   - Do **not** infer selection criteria from artifact structure (e.g., do not treat “meeting participants” as “reviewers” unless the question says so).\n2. **Build a candidate artifact set (wide recall, then filter):**\n   - Start with product-scoped containers when available (e.g., `\u002Froot\u002FDATA\u002Fproducts\u002F\u003CProduct>.json`).\n   - Expand to global search only as needed (grep\u002Fripgrep over `\u002Froot\u002FDATA` for exact strings from the question: product name, report title, competitor name, URL fragments, doc IDs).\n   - Follow references across artifacts (slack ↔ doc links ↔ meeting transcripts ↔ PRs) to reach primary sources.\n3. **HARD grounding & anti-distractor gate (prevent cross-product leakage):**\n   - Accept a candidate artifact only if it passes **2+ independent grounding signals** relevant to the question’s target scope, such as:\n     - Located under the correct product container *and* referenced from product-specific channels\u002Fmeetings.\n     - Title\u002Fcontent explicitly mentions the target product\u002Ftopic.\n     - Appears in a clearly product-scoped channel\u002Fmeeting series.\n     - Link\u002Fpath includes a product-specific identifier consistent with the target product.\n   - Reject artifacts that primarily concern a different product\u002Ftopic unless the question explicitly requests cross-product comparisons.\n4. **Select the correct “version” when multiple candidates exist:**\n   - Prefer explicit `latest`\u002F`final` markers; else use most recent date; else use the one most referenced in follow-up discussions tied to the same grounded scope.\n5. **Extract entities with evidence-based rules (no role assumptions):**\n   - Authors: prefer explicit fields (`author\u002Fauthors\u002Fcreated_by\u002Fowner`), then PR author, then the Slack poster *only if* it clearly introduces the exact artifact.\n   - Reviewers\u002Fcontributors: use explicit reviewer\u002Fapprover fields first; else include only people who provide **substantive** feedback in doc comments, meeting transcript turns, or Slack thread replies that are clearly tied to the artifact.\n   - URLs\u002Fitems\u002FIDs: extract directly from artifact fields or message text; de-duplicate while preserving order.\n6. **Normalize + validate outputs:**\n   - Normalize to the requested type (e.g., employee IDs as `eid_...`).\n   - Validate every returned employee ID exists in `\u002Froot\u002FDATA\u002Fmetadata\u002Femployee.json` (or the dataset’s employee directory).\n   - Ensure the final answer for each question is **always a list** (even length-1).\n7. **Write results in the required JSON shape:**\n   - Write `\u002Froot\u002Fanswer.json` as:\n     - top-level dict keyed by question key (`\"q1\"`, `\"q2\"`, …),\n     - each value is `{\"answer\": [...], \"tokens\": \"...\"}`.\n   - Keep `tokens` as a **string**, matching the task’s example schema.\n8. **Token logging (avoid failing installs; prefer available tooling; be explicit if approximate):**\n   - Do **not** attempt to `pip install` tokenizers in restricted environments.\n   - First try to compute token counts using already-available libraries (e.g., `python -c \"import tiktoken\"`). If available, use it to count tokens for the per-question serialized payload you produce.\n   - If no tokenizer library is available, record a transparent string value such as `\"approx:\u003Cn>\"` based on a consistent local heuristic (e.g., character-count\u002F4), rather than implying an exact “consumed tokens” measurement.\n   - Never store numeric token values; always store the token field as a string.\n## Expected Result\nFor every question in `\u002Froot\u002Fquestion.txt`, `\u002Froot\u002Fanswer.json` contains an entry like:\n- `\"answer\"` is a list (deduplicated, validated, grounded in artifacts),\n- `\"tokens\"` is a **string** computed via an available tokenizer when possible, otherwise transparently marked as approximate (e.g., `\"approx:123\"`),\nand entity selection is driven by explicit question requirements and direct artifact evidence (not assumptions like “participants == reviewers”).\n","## Steps\n1. **Ground on the question text (no heuristics):**\n   - Read `\u002Froot\u002Fquestion.txt` and treat each entry (e.g., `q1`, `q2`, …) as the *only* specification for what to retrieve.\n   - For each question, explicitly extract: target product\u002Ftopic, required entity types (IDs\u002Fnames\u002FURLs\u002Fetc.), and any constraints (“author”, “key reviewers”, “competitor demos”, “from Slack”, “latest”, etc.).\n   - Do **not** infer selection criteria from artifact structure (e.g., do not treat “meeting participants” as “reviewers” unless the question says so).\n2. **Build a candidate artifact set (wide recall, then filter):**\n   - Start with product-scoped containers when available (e.g., `\u002Froot\u002FDATA\u002Fproducts\u002F\u003CProduct>.json`).\n   - Expand to global search only as needed (grep\u002Fripgrep over `\u002Froot\u002FDATA` for exact strings from the question: product name, report title, competitor name, URL fragments, doc IDs).\n   - Follow references across artifacts (slack ↔ doc links ↔ meeting transcripts ↔ PRs) to reach primary sources.\n3. **HARD grounding & anti-distractor gate (prevent cross-product leakage):**\n   - Accept a candidate artifact only if it passes **2+ independent grounding signals** relevant to the question’s target scope, such as:\n     - Located under the correct product container *and* referenced from product-specific channels\u002Fmeetings.\n     - Title\u002Fcontent explicitly mentions the target product\u002Ftopic.\n     - Appears in a clearly product-scoped channel\u002Fmeeting series.\n     - Link\u002Fpath includes a product-specific identifier consistent with the target product.\n   - Reject artifacts that primarily concern a different product\u002Ftopic unless the question explicitly requests cross-product comparisons.\n4. **Select the correct “version” when multiple candidates exist:**\n   - Prefer explicit `latest`\u002F`final` markers; else use most recent date; else use the one most referenced in follow-up discussions tied to the same grounded scope.\n5. **Extract entities with evidence-based rules (no role assumptions):**\n   - Authors: prefer explicit fields (`author\u002Fauthors\u002Fcreated_by\u002Fowner`), then PR author, then the Slack poster *only if* it clearly introduces the exact artifact.\n   - Reviewers\u002Fcontributors: use explicit reviewer\u002Fapprover fields first; else include only people who provide **substantive** feedback in doc comments, meeting transcript turns, or Slack thread replies that are clearly tied to the artifact.\n   - URLs\u002Fitems\u002FIDs: extract directly from artifact fields or message text; de-duplicate while preserving order.\n6. **Normalize + validate outputs:**\n   - Normalize to the requested type (e.g., employee IDs as `eid_...`).\n   - Validate every returned employee ID exists in `\u002Froot\u002FDATA\u002Fmetadata\u002Femployee.json` (or the dataset’s employee directory).\n   - Ensure the final answer for each question is **always a list** (even length-1).\n7. **Write results in the required JSON shape:**\n   - Write `\u002Froot\u002Fanswer.json` as:\n     - top-level dict keyed by question key (`\"q1\"`, `\"q2\"`, …),\n     - each value is `{\"answer\": [...], \"tokens\": \"...\"}`.\n   - Keep `tokens` as a **string**, matching the task’s example schema.\n8. **Token logging (avoid failing installs; prefer available tooling; be explicit if approximate):**\n   - Do **not** attempt to `pip install` tokenizers in restricted environments.\n   - First try to compute token counts using already-available libraries (e.g., `python -c \"import tiktoken\"`). If available, use it to count tokens for the per-question serialized payload you produce.\n   - If no tokenizer library is available, record a transparent string value such as `\"approx:\u003Cn>\"` based on a consistent local heuristic (e.g., character-count\u002F4), rather than implying an exact “consumed tokens” measurement.\n   - Never store numeric token values; always store the token field as a string.\n## Expected Result\nFor every question in `\u002Froot\u002Fquestion.txt`, `\u002Froot\u002Fanswer.json` contains an entry like:\n- `\"answer\"` is a list (deduplicated, validated, grounded in artifacts),\n- `\"tokens\"` is a **string** computed via an available tokenizer when possible, otherwise transparently marked as approximate (e.g., `\"approx:123\"`),\nand entity selection is driven by explicit question requirements and direct artifact evidence (not assumptions like “participants == reviewers”).\n",{"name":5,"description":7},"main","cdcb6d230f3b15c3c33f8a2bdd16439f43602cbe","evaluation\u002Fresults\u002FMMG2Skill-GPT52-Gold31-20260923\u002Ftasks\u002Fenterprise-information-search\u002Frepaired_skill\u002Fenterprise-artifact-search",1,"https:\u002F\u002Fgithub.com\u002Flinyouxue\u002Fskill-repair-benchmark","https:\u002F\u002Fgithub.com\u002Flinyouxue","2026-11-14 20:28:35",{"source":8,"skillName":5,"commands":32,"prompt":42,"zipAvailable":43},[33,36,39],{"tool":34,"command":35},"npx","npx skills add linyouxue\u002Fskill-repair-benchmark --skill enterprise-artifact-search",{"tool":37,"command":38},"bunx","bunx skills add linyouxue\u002Fskill-repair-benchmark --skill enterprise-artifact-search",{"tool":40,"command":41},"pnpm","pnpm dlx skills add linyouxue\u002Fskill-repair-benchmark --skill enterprise-artifact-search","安装技能，技能地址：https:\u002F\u002Fgithub.com\u002Flinyouxue\u002Fskill-repair-benchmark\u002Fblob\u002Fmain\u002Fevaluation\u002Fresults\u002FMMG2Skill-GPT52-Gold31-20260923\u002Ftasks\u002Fenterprise-information-search\u002Frepaired_skill\u002Fenterprise-artifact-search\u002FSKILL.md",true,[],[46,53,60,67,72,77],{"skillNo":47,"name":48,"nameZh":6,"description":49,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":50,"categories":51,"keywords":52,"publishedAt":19},"01M7E1HGVRY73NBJ022CJG13X6","video-frame-extraction","Extract frames from video files and save them as images using OpenCV","https:\u002F\u002Fgithub.com\u002Flinyouxue\u002Fskill-repair-benchmark\u002Fblob\u002Fmain\u002Fjpg-ocr-stat\u002Fresult\u002Fskilltta-claude4.7\u002Fgroup_03\u002Frun_03\u002Finputs\u002Fskills\u002Fvideo-frame-extraction\u002FSKILL.md",[],[],{"skillNo":54,"name":55,"nameZh":6,"description":56,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":57,"categories":58,"keywords":59,"publishedAt":19},"01M7E1HAG5R7CJ9C8NYJS771MB","openai-vision","Analyze images and multi-frame sequences using OpenAI GPT vision models","https:\u002F\u002Fgithub.com\u002Flinyouxue\u002Fskill-repair-benchmark\u002Fblob\u002Fmain\u002Fjpg-ocr-stat\u002Fresult\u002Fskilltta-claude4.7\u002Fgroup_03\u002Frun_03\u002Finputs\u002Fskills\u002Fopenai-vision\u002FSKILL.md",[],[],{"skillNo":61,"name":62,"nameZh":6,"description":63,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":64,"categories":65,"keywords":66,"publishedAt":19},"01M7E1H7DADKB6QBWKC2KQ63SZ","image-ocr","Extract text content from images using Tesseract OCR via Python, including structured field extraction from receipts and forms","https:\u002F\u002Fgithub.com\u002Flinyouxue\u002Fskill-repair-benchmark\u002Fblob\u002Fmain\u002Fjpg-ocr-stat\u002Fresult\u002Fskilltta-claude4.7\u002Fgroup_03\u002Frun_03\u002Finputs\u002Fskills\u002Fimage-ocr\u002FSKILL.md",[],[],{"skillNo":68,"name":48,"nameZh":6,"description":49,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":69,"categories":70,"keywords":71,"publishedAt":19},"01M7E1H0WBNT43TENGQY7AVZHD","https:\u002F\u002Fgithub.com\u002Flinyouxue\u002Fskill-repair-benchmark\u002Fblob\u002Fmain\u002Fjpg-ocr-stat\u002Fresult\u002Fskilltta-claude4.7\u002Fgroup_03\u002Frun_02\u002Finputs\u002Fskills\u002Fvideo-frame-extraction\u002FSKILL.md",[],[],{"skillNo":73,"name":55,"nameZh":6,"description":56,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":74,"categories":75,"keywords":76,"publishedAt":19},"01M7E1GTC7D4N4Q2EF03875J9E","https:\u002F\u002Fgithub.com\u002Flinyouxue\u002Fskill-repair-benchmark\u002Fblob\u002Fmain\u002Fjpg-ocr-stat\u002Fresult\u002Fskilltta-claude4.7\u002Fgroup_03\u002Frun_02\u002Finputs\u002Fskills\u002Fopenai-vision\u002FSKILL.md",[],[],{"skillNo":78,"name":62,"nameZh":6,"description":63,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":79,"categories":80,"keywords":81,"publishedAt":19},"01M7E1GQ92CWFMGH53KSSE4WER","https:\u002F\u002Fgithub.com\u002Flinyouxue\u002Fskill-repair-benchmark\u002Fblob\u002Fmain\u002Fjpg-ocr-stat\u002Fresult\u002Fskilltta-claude4.7\u002Fgroup_03\u002Frun_02\u002Finputs\u002Fskills\u002Fimage-ocr\u002FSKILL.md",[],[]]