[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"skill:01M7C1RKQEEJCKTC3BD9E8F9GF":3,"skill-same:01M7C1RKQEEJCKTC3BD9E8F9GF":43},{"skillNo":4,"name":5,"nameZh":6,"description":6,"descriptionZh":6,"repositoryFullName":7,"ownerLogin":8,"ownerAvatarUrl":9,"ownerAccountNo":10,"ownerDisplayName":11,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":15,"categories":16,"keywords":17,"publishedAt":18,"linkOnly":19,"markdownContent":20,"markdownBody":20,"frontMatter":21,"refName":22,"commitSha":23,"skillDirectory":24,"forksCount":25,"repositoryHtmlUrl":26,"repositoryHomepageUrl":6,"ownerHtmlUrl":27,"ownerBlogUrl":6,"ownerDescription":6,"updatedAt":28,"install":29,"resources":42},"01M7C1RKQEEJCKTC3BD9E8F9GF","enterprise-artifact-search",null,"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\u002FSkillrevise\u002Ftasks\u002Fenterprise-information-search\u002Frepaired_skill\u002Fenterprise-artifact-search\u002FSKILL.md",[],[],"2026-11-15 14:19:17",false,"# QA File Loop To Verifier-Aligned answer.json\n\n## Purpose\nProduce verifier-aligned answers for multiple questions stored in a local question file by looping over discovered question ids, answering each using minimal necessary evidence lookup, and writing a single JSON artifact to the required output path with a schema-validated structure.\n\n## When to Use\nUse when a task requires: (a) reading one local question file (commonly \u002Froot\u002Fquestion.txt), (b) generating answers for multiple keyed questions, and (c) writing a machine-checked JSON file (commonly \u002Froot\u002Fanswer.json). Do not use for single-question, conversational answers that do not require an on-disk JSON artifact.\n\n## Procedure\n- Step 1: Discover inputs and required outputs (no assumptions)\n- Confirm the question file path from the task instructions; if not provided, search minimally (current directory, then \u002Froot) for a likely question file and stop to ask if ambiguous.\n- Open and parse the question file to extract a stable list of (question_id, question_text).\n- Decision point: If ids are missing\u002Fambiguous (for example, free text without clear separators), stop and request clarification or define a deterministic parsing rule and document it in the output.\n- Step 2: For each question id, answer with bounded, evidence-first lookup\n- For each (question_id, question_text):\n- Identify what must be looked up vs. derived (names, ids, dates, counts, file contents).\n- Perform targeted discovery in the local environment (list relevant directories, grep\u002Fsearch for exact identifiers) before opening large artifacts.\n- Extract only what is needed to answer; keep a short internal evidence note (file name + line\u002Frecord pointer) so you can re-check quickly.\n- Decision point: If multiple plausible answers remain, choose the one best supported by direct artifact text; otherwise mark the answer as ambiguous in the answer field rather than guessing.\n- Step 3: Populate consumed_tokens with a contract-safe rule (no environment mutation)\n- Do not install packages (no pip\u002Fapt\u002Fvenv) solely to compute tokens.\n- If the runtime\u002Ftooling provides an explicit token-usage metric for the specific question (for example, per-call usage from the platform), record that integer.\n- Otherwise, set consumed_tokens to null (or omit only if the task explicitly allows omission). Do not fabricate heuristic estimates and present them as true consumption.\n- Step 4: Write \u002Froot\u002Fanswer.json and validate it matches the discovered contract (final checkpoint)\n- Build an object mapping each discovered question_id to an object containing:\n- answer: the per-question answer (string or JSON-serializable object as required by the task)\n- consumed_tokens: integer usage if available, else null per Step 3\n- Write the JSON to the exact required output path (commonly \u002Froot\u002Fanswer.json). Do not write to an alternate path.\n- Execution anchor (must pass before finishing): reload \u002Froot\u002Fanswer.json and validate:\n- File exists and is readable at the required path.\n- JSON parses successfully.\n- Top-level keys exactly match the discovered question ids (no missing, no extras).\n- Each value is an object containing answer and consumed_tokens with acceptable types (answer is JSON-serializable; consumed_tokens is integer or null per Step 3).\n- If validation fails, fix the construction and rewrite, then re-validate.\n\n## Constraints \u002F Pitfalls\n- Do not mutate the environment to satisfy bookkeeping (no pip\u002Fapt installs for token counting). If token usage is not available from the platform, use a null\u002Funknown policy rather than a heuristic number.\n- Do not over-scan or load large artifacts by default: start with discovery (directory listing, targeted search) and cap retrieval to what the question requires; if requirements are ambiguous, stop and ask rather than guessing.",{},"main","cdcb6d230f3b15c3c33f8a2bdd16439f43602cbe","evaluation\u002Fresults\u002FSkillrevise\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":7,"skillName":5,"commands":30,"prompt":40,"zipAvailable":41},[31,34,37],{"tool":32,"command":33},"npx","npx skills add linyouxue\u002Fskill-repair-benchmark --skill enterprise-artifact-search",{"tool":35,"command":36},"bunx","bunx skills add linyouxue\u002Fskill-repair-benchmark --skill enterprise-artifact-search",{"tool":38,"command":39},"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\u002FSkillrevise\u002Ftasks\u002Fenterprise-information-search\u002Frepaired_skill\u002Fenterprise-artifact-search\u002FSKILL.md",true,[],[44,51,58,65,70,75],{"skillNo":45,"name":46,"nameZh":6,"description":47,"descriptionZh":6,"repositoryFullName":7,"ownerLogin":8,"ownerAvatarUrl":9,"ownerAccountNo":10,"ownerDisplayName":11,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":48,"categories":49,"keywords":50,"publishedAt":18},"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":52,"name":53,"nameZh":6,"description":54,"descriptionZh":6,"repositoryFullName":7,"ownerLogin":8,"ownerAvatarUrl":9,"ownerAccountNo":10,"ownerDisplayName":11,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":55,"categories":56,"keywords":57,"publishedAt":18},"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":59,"name":60,"nameZh":6,"description":61,"descriptionZh":6,"repositoryFullName":7,"ownerLogin":8,"ownerAvatarUrl":9,"ownerAccountNo":10,"ownerDisplayName":11,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":62,"categories":63,"keywords":64,"publishedAt":18},"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":66,"name":46,"nameZh":6,"description":47,"descriptionZh":6,"repositoryFullName":7,"ownerLogin":8,"ownerAvatarUrl":9,"ownerAccountNo":10,"ownerDisplayName":11,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":67,"categories":68,"keywords":69,"publishedAt":18},"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":71,"name":53,"nameZh":6,"description":54,"descriptionZh":6,"repositoryFullName":7,"ownerLogin":8,"ownerAvatarUrl":9,"ownerAccountNo":10,"ownerDisplayName":11,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":72,"categories":73,"keywords":74,"publishedAt":18},"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":76,"name":60,"nameZh":6,"description":61,"descriptionZh":6,"repositoryFullName":7,"ownerLogin":8,"ownerAvatarUrl":9,"ownerAccountNo":10,"ownerDisplayName":11,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":77,"categories":78,"keywords":79,"publishedAt":18},"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",[],[]]