[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"skill:01M6ADC8EBNYYN6XNERHR5YYMJ":3,"skill-same:01M6ADC8EBNYYN6XNERHR5YYMJ":96},{"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":29,"ownerHtmlUrl":30,"ownerBlogUrl":31,"ownerDescription":32,"updatedAt":33,"install":34,"resources":47},"01M6ADC8EBNYYN6XNERHR5YYMJ","localizations",null,"A skill that localizes contents into given locales.","github\u002Fcopilot-sdk-workshop","github","https:\u002F\u002Favatars.githubusercontent.com\u002Fu\u002F9919?v=4","01M6AC83DS133TPXFSP1DHNCC6","GitHub","01M6ADC6EEHG4PB90B1S0WP8YS",8,"MIT","https:\u002F\u002Fgithub.com\u002Fgithub\u002Fcopilot-sdk-workshop\u002Fblob\u002Fmain\u002F.github\u002Fskills\u002Flocalizations\u002FSKILL.md",[],[5],"2026-11-10 10:12:30",false,"---\nname: eval-suite-planner\ndescription: Plan standalone — populates the Eval Suite Planning & Logging Template from an Agent Vision or plain-English agent description. Grounded in Practical Guidance on Agent Evaluation v5: Step 1 planning, Steps 2-3 eval-set decomposition, Step 4 gates\u002Fimprovement targets, Step 5 human inputs, Step 6 grader-validation planning, Step 7 baseline placeholders, Step 8 regression partitioning, and Step 10 reusable-asset candidates. Output is a template-preserving `.xlsx` workbook plus an interactive HTML review page. Use before generating test cases or running evals.\n---\n\n## Purpose\n\nThis skill produces the **Plan** artifact of the `\u002Feval-guide` lifecycle: a populated copy of the customer's **Eval Suite Planning & Logging Template** plus an interactive HTML review page. The workbook is the source-of-truth artifact; do not replace it with a scenario table, quality-signal table, generic spreadsheet, default `.docx` report, or HTML-only plan.\n\nThe skill aligns to `skills\u002Feval-guide\u002Fplaybook.md` and `skills\u002Feval-guide\u002Feval-suite-template.md`. Use the 10-step playbook as the methodology spine and the XLSX template as the output shape.\n\n## Core rule\n\n**Copy the blank XLSX template and populate existing cells\u002Frows only. Do not modify the template.**\n\nDo not rename sheets, add sheets, delete sheets, add columns, change headers, rewrite README text, edit `Dropdown Lists`, change styles, change data validation, or convert the template into a different spreadsheet.\n\nIf a blank template workbook is available in the session, use it. If not, ask the user to provide the template; do not silently invent a new workbook.\n\n## Question policy\n\nAsk targeted questions only when a workbook field materially affects the plan and cannot be inferred safely:\n\n1. Eval owner \u002F named approver.\n2. Lifecycle stage and target deployment decision.\n3. Whether the agent is prompt-only, RAG\u002Fknowledge-grounded, or agentic with tools\u002Fconnectors.\n4. Regulated\u002Fcompliance obligations.\n5. Authoritative sources and source owners.\n\nIf the user wants speed or cannot answer, populate `TBD - confirm before baseline`.\n\n## Planning method\n\nWhen invoked as `\u002Feval-suite-planner \u003Cagent description>`:\n\n1. Extract or infer the agent's purpose, users, knowledge sources, capabilities, boundaries, architecture, lifecycle stage, and known risks.\n2. Populate **Step 1 — Plan the Eval Effort**:\n   - one-sentence eval objective;\n   - five-factor risk tier: reach, criticality of error, autonomy\u002Fblast radius, regulatory\u002Fcompliance exposure, data sensitivity;\n   - one accountable owner.\n3. Define eval sets, not scenarios:\n   - **Capability eval sets**: one row per capability dimension that must be diagnostic, e.g. accuracy\u002Fcorrectness, faithfulness\u002Fgroundedness, relevancy, style\u002Ftone, reasoning\u002Ftool use.\n   - **Trust & Safety eval sets**: one row per refusal, boundary, or safety category, e.g. guardrails, out-of-scope handling, sensitive-data handling, prompt injection\u002Fjailbreak, compliance-specific behavior.\n4. Apply **Step 4 v5 gates\u002Fimprovement-target logic**:\n   - T&S sets use absolute pass-rate hard gates, usually near 100%.\n   - Capability sets usually use a launch floor for first deployment plus regression\u002Fdirection after baseline, not a standing absolute pass-rate target.\n   - High-risk capabilities that function like guardrails keep explicit hard floors.\n   - Use the template's existing `Target pass rate`, `Target rationale`, `Gate type`, `Intended use`, `Run cadence`, and `Notes` columns to express this; do not add a new column.\n5. Specify **Step 5 human inputs**:\n   - grading rubric, ground truth, golden answer, or rubric + ground truth;\n   - author\u002Fowner;\n   - grounding source dependency;\n   - whether source changes require review.\n6. Plan **Step 6 grader validation** without changing the template:\n   - record grader type and validation expectation in the registry row's `Notes`;\n   - for LLM-as-judge \u002F Custom rubrics, note that human-labeled hard and borderline cases must validate the judge before baseline scores are trusted;\n   - for programmatic checks, note the deterministic check to confirm;\n   - for human grading, note reviewer agreement expectations where relevant.\n7. Seed **Step 7 baseline placeholders** in `3 . Run Log` only when useful:\n   - one placeholder row per eval set;\n   - `Run type = Baseline`;\n   - result fields blank;\n   - `Actionable next step = Validate grader, then run baseline`;\n   - `Status = Open`.\n8. Apply **Step 8 regression partitioning** in existing registry fields:\n   - capability sets usually `Intended use = Both` or `Regression`;\n   - most T&S sets are `Gate`; the slim subset likely affected by model\u002Ftool\u002Fpolicy changes can be `Both` or `Regression`;\n   - set `Run cadence` using existing dropdown values such as `Per-change`, `Nightly`, `Weekly`, or `Milestone-only`.\n9. Flag **Step 10 reusable assets** in `4 . Reusable Library`:\n   - reusable T&S sets;\n   - grading rubrics;\n   - failure-pattern templates;\n   - production-derived edge-case categories when applicable.\n\n## Workbook population rules\n\nUse `skills\u002Feval-guide\u002Feval-suite-template.md` as the exact tab\u002Fcolumn map.\n\n### `README`\n\nDo not edit.\n\n### `1 . Planning`\n\nPopulate only existing input cells:\n\n- Agent identity.\n- Risk classification (5 factors).\n- Owners & roles.\n- Deployment gates \u002F sign-off criteria.\n\nFor the template's `Min pass rate - Capability` row, reflect v5 Step 4 accurately: use launch floor \u002F high-risk capability floor \u002F regression-governance language, not a generic scenario pass-rate target.\n\n### `2 . Eval Suite Registry`\n\nPopulate one row per eval set. Do **not** populate one row per test case or legacy planning artifact.\n\nRequired row semantics:\n\n- `Category`: `Capability` or `Trust & Safety`.\n- `Dimension tested`: capability dimension or T&S category from the template dropdowns.\n- `Purpose \u002F diagnostic signal`: what failure in this set diagnoses.\n- `Target pass rate`: absolute gate for T&S; launch floor or `Regression \u002F direction after baseline` for most capability sets.\n- `Target rationale`: v5 Step 4 rationale.\n- `Gate type`: closest existing dropdown value.\n- `Intended use`: `Gate`, `Regression`, or `Both`.\n- `Run cadence`: cadence for Step 8.\n- `Human input type`, `Human input author`, `Grounding source dependency`, `Source change -> review?`: Step 5.\n- `Reusable asset?`, `Reuse tier`, `Set status`: Step 10 and lifecycle status.\n- `Notes`: assumptions, open questions, Step 4 nuance, and Step 6 grader-validation plan.\n\n### `3 . Run Log`\n\nUse this for Step 7 baseline\u002Fiteration logging. During planning, add placeholder baseline rows only if useful; keep result fields blank.\n\n### `4 . Reusable Library`\n\nPopulate candidate reusable assets only. Do not duplicate every eval set; promote assets that could help other agents.\n\n### `Dropdown Lists`\n\nDo not edit.\n\n## Output\n\nCreate `eval-suite-\u003Cagent-name>-\u003CYYYY-MM-DD>.xlsx` as a populated copy of the template.\n\nThen create `eval-suite-\u003Cagent-name>-\u003CYYYY-MM-DD>-review.html` next to the workbook using `skills\u002Feval-guide\u002Fplan-review-page.md`.\n\nDo not paste the summary, eval-set table, or checklist into chat. The HTML page carries that content. The final chat response should be only the workbook path, the HTML review page path, and any blocker\u002Fmanual action.\n\n## Human review checkpoints\n\nInclude these in the HTML review page checklist instead of displaying them in chat:\n\n| # | Checkpoint | What to verify |\n|---|---|---|\n| 1 | Objective, risk tier, owner | The objective is decision-oriented, the five-factor risk tier is right, and a named owner can sign off. |\n| 2 | Eval-set decomposition | Capability sets isolate one diagnostic capability each; T&S sets remain separate from capability. |\n| 3 | Step 4 bars | T&S has absolute hard gates; capability uses launch floors \u002F regression-direction unless high-risk. |\n| 4 | Human inputs | Rubrics, ground truths, golden answers, and source dependencies have owners. |\n| 5 | Grader validation | Each set has a plausible grader type and validation plan before baseline. |\n| 6 | Regression partition | Capability and slim T&S regression sets have cadence; gate-only T&S sets run at milestones. |\n| 7 | Template integrity | No sheets, columns, headers, dropdowns, README text, or formatting were changed. |\n\n## Behavior rules\n\n- Do not generate scenario-plan tables as the Plan artifact.\n- Do not generate quality-signal sheets or quality-signal grouping as the Plan artifact.\n- Do not add columns to support missing concepts; use existing fields, especially `Notes`.\n- Do not create a `.docx` unless the user explicitly asks for a narrative report.\n- Do not produce long narrative chat output after artifact generation; use the HTML review page for the interactive summary and checkpoints.\n- Be specific to the described agent, but at eval-set granularity.\n\n## Companion skills\n\n- **`\u002Feval-generator`** — Generate test cases from the populated workbook registry.\n- **`\u002Feval-result-interpreter`** — Interpret baseline \u002F iteration results using Step 6-7 and gate status.\n- **`\u002Feval-triage-and-improvement`** — Diagnose failures and feed the Step 9 optimization loop.\n- **`\u002Feval-library-promoter`** — Promote Step 10 reusable assets.\n- **`\u002Feval-guide`** — Orchestrated workflow with dashboard review checkpoints.\n","## Purpose\n\nThis skill produces the **Plan** artifact of the `\u002Feval-guide` lifecycle: a populated copy of the customer's **Eval Suite Planning & Logging Template** plus an interactive HTML review page. The workbook is the source-of-truth artifact; do not replace it with a scenario table, quality-signal table, generic spreadsheet, default `.docx` report, or HTML-only plan.\n\nThe skill aligns to `skills\u002Feval-guide\u002Fplaybook.md` and `skills\u002Feval-guide\u002Feval-suite-template.md`. Use the 10-step playbook as the methodology spine and the XLSX template as the output shape.\n\n## Core rule\n\n**Copy the blank XLSX template and populate existing cells\u002Frows only. Do not modify the template.**\n\nDo not rename sheets, add sheets, delete sheets, add columns, change headers, rewrite README text, edit `Dropdown Lists`, change styles, change data validation, or convert the template into a different spreadsheet.\n\nIf a blank template workbook is available in the session, use it. If not, ask the user to provide the template; do not silently invent a new workbook.\n\n## Question policy\n\nAsk targeted questions only when a workbook field materially affects the plan and cannot be inferred safely:\n\n1. Eval owner \u002F named approver.\n2. Lifecycle stage and target deployment decision.\n3. Whether the agent is prompt-only, RAG\u002Fknowledge-grounded, or agentic with tools\u002Fconnectors.\n4. Regulated\u002Fcompliance obligations.\n5. Authoritative sources and source owners.\n\nIf the user wants speed or cannot answer, populate `TBD - confirm before baseline`.\n\n## Planning method\n\nWhen invoked as `\u002Feval-suite-planner \u003Cagent description>`:\n\n1. Extract or infer the agent's purpose, users, knowledge sources, capabilities, boundaries, architecture, lifecycle stage, and known risks.\n2. Populate **Step 1 — Plan the Eval Effort**:\n   - one-sentence eval objective;\n   - five-factor risk tier: reach, criticality of error, autonomy\u002Fblast radius, regulatory\u002Fcompliance exposure, data sensitivity;\n   - one accountable owner.\n3. Define eval sets, not scenarios:\n   - **Capability eval sets**: one row per capability dimension that must be diagnostic, e.g. accuracy\u002Fcorrectness, faithfulness\u002Fgroundedness, relevancy, style\u002Ftone, reasoning\u002Ftool use.\n   - **Trust & Safety eval sets**: one row per refusal, boundary, or safety category, e.g. guardrails, out-of-scope handling, sensitive-data handling, prompt injection\u002Fjailbreak, compliance-specific behavior.\n4. Apply **Step 4 v5 gates\u002Fimprovement-target logic**:\n   - T&S sets use absolute pass-rate hard gates, usually near 100%.\n   - Capability sets usually use a launch floor for first deployment plus regression\u002Fdirection after baseline, not a standing absolute pass-rate target.\n   - High-risk capabilities that function like guardrails keep explicit hard floors.\n   - Use the template's existing `Target pass rate`, `Target rationale`, `Gate type`, `Intended use`, `Run cadence`, and `Notes` columns to express this; do not add a new column.\n5. Specify **Step 5 human inputs**:\n   - grading rubric, ground truth, golden answer, or rubric + ground truth;\n   - author\u002Fowner;\n   - grounding source dependency;\n   - whether source changes require review.\n6. Plan **Step 6 grader validation** without changing the template:\n   - record grader type and validation expectation in the registry row's `Notes`;\n   - for LLM-as-judge \u002F Custom rubrics, note that human-labeled hard and borderline cases must validate the judge before baseline scores are trusted;\n   - for programmatic checks, note the deterministic check to confirm;\n   - for human grading, note reviewer agreement expectations where relevant.\n7. Seed **Step 7 baseline placeholders** in `3 . Run Log` only when useful:\n   - one placeholder row per eval set;\n   - `Run type = Baseline`;\n   - result fields blank;\n   - `Actionable next step = Validate grader, then run baseline`;\n   - `Status = Open`.\n8. Apply **Step 8 regression partitioning** in existing registry fields:\n   - capability sets usually `Intended use = Both` or `Regression`;\n   - most T&S sets are `Gate`; the slim subset likely affected by model\u002Ftool\u002Fpolicy changes can be `Both` or `Regression`;\n   - set `Run cadence` using existing dropdown values such as `Per-change`, `Nightly`, `Weekly`, or `Milestone-only`.\n9. Flag **Step 10 reusable assets** in `4 . Reusable Library`:\n   - reusable T&S sets;\n   - grading rubrics;\n   - failure-pattern templates;\n   - production-derived edge-case categories when applicable.\n\n## Workbook population rules\n\nUse `skills\u002Feval-guide\u002Feval-suite-template.md` as the exact tab\u002Fcolumn map.\n\n### `README`\n\nDo not edit.\n\n### `1 . Planning`\n\nPopulate only existing input cells:\n\n- Agent identity.\n- Risk classification (5 factors).\n- Owners & roles.\n- Deployment gates \u002F sign-off criteria.\n\nFor the template's `Min pass rate - Capability` row, reflect v5 Step 4 accurately: use launch floor \u002F high-risk capability floor \u002F regression-governance language, not a generic scenario pass-rate target.\n\n### `2 . Eval Suite Registry`\n\nPopulate one row per eval set. Do **not** populate one row per test case or legacy planning artifact.\n\nRequired row semantics:\n\n- `Category`: `Capability` or `Trust & Safety`.\n- `Dimension tested`: capability dimension or T&S category from the template dropdowns.\n- `Purpose \u002F diagnostic signal`: what failure in this set diagnoses.\n- `Target pass rate`: absolute gate for T&S; launch floor or `Regression \u002F direction after baseline` for most capability sets.\n- `Target rationale`: v5 Step 4 rationale.\n- `Gate type`: closest existing dropdown value.\n- `Intended use`: `Gate`, `Regression`, or `Both`.\n- `Run cadence`: cadence for Step 8.\n- `Human input type`, `Human input author`, `Grounding source dependency`, `Source change -> review?`: Step 5.\n- `Reusable asset?`, `Reuse tier`, `Set status`: Step 10 and lifecycle status.\n- `Notes`: assumptions, open questions, Step 4 nuance, and Step 6 grader-validation plan.\n\n### `3 . Run Log`\n\nUse this for Step 7 baseline\u002Fiteration logging. During planning, add placeholder baseline rows only if useful; keep result fields blank.\n\n### `4 . Reusable Library`\n\nPopulate candidate reusable assets only. Do not duplicate every eval set; promote assets that could help other agents.\n\n### `Dropdown Lists`\n\nDo not edit.\n\n## Output\n\nCreate `eval-suite-\u003Cagent-name>-\u003CYYYY-MM-DD>.xlsx` as a populated copy of the template.\n\nThen create `eval-suite-\u003Cagent-name>-\u003CYYYY-MM-DD>-review.html` next to the workbook using `skills\u002Feval-guide\u002Fplan-review-page.md`.\n\nDo not paste the summary, eval-set table, or checklist into chat. The HTML page carries that content. The final chat response should be only the workbook path, the HTML review page path, and any blocker\u002Fmanual action.\n\n## Human review checkpoints\n\nInclude these in the HTML review page checklist instead of displaying them in chat:\n\n| # | Checkpoint | What to verify |\n|---|---|---|\n| 1 | Objective, risk tier, owner | The objective is decision-oriented, the five-factor risk tier is right, and a named owner can sign off. |\n| 2 | Eval-set decomposition | Capability sets isolate one diagnostic capability each; T&S sets remain separate from capability. |\n| 3 | Step 4 bars | T&S has absolute hard gates; capability uses launch floors \u002F regression-direction unless high-risk. |\n| 4 | Human inputs | Rubrics, ground truths, golden answers, and source dependencies have owners. |\n| 5 | Grader validation | Each set has a plausible grader type and validation plan before baseline. |\n| 6 | Regression partition | Capability and slim T&S regression sets have cadence; gate-only T&S sets run at milestones. |\n| 7 | Template integrity | No sheets, columns, headers, dropdowns, README text, or formatting were changed. |\n\n## Behavior rules\n\n- Do not generate scenario-plan tables as the Plan artifact.\n- Do not generate quality-signal sheets or quality-signal grouping as the Plan artifact.\n- Do not add columns to support missing concepts; use existing fields, especially `Notes`.\n- Do not create a `.docx` unless the user explicitly asks for a narrative report.\n- Do not produce long narrative chat output after artifact generation; use the HTML review page for the interactive summary and checkpoints.\n- Be specific to the described agent, but at eval-set granularity.\n\n## Companion skills\n\n- **`\u002Feval-generator`** — Generate test cases from the populated workbook registry.\n- **`\u002Feval-result-interpreter`** — Interpret baseline \u002F iteration results using Step 6-7 and gate status.\n- **`\u002Feval-triage-and-improvement`** — Diagnose failures and feed the Step 9 optimization loop.\n- **`\u002Feval-library-promoter`** — Promote Step 10 reusable assets.\n- **`\u002Feval-guide`** — Orchestrated workflow with dashboard review checkpoints.\n",{},"main","7a22a8957ced79a35fd7b4634fd5b5eac56c308e",".github\u002Fskills\u002Flocalizations",28,"https:\u002F\u002Fgithub.com\u002Fgithub\u002Fcopilot-sdk-workshop","https:\u002F\u002Fgithub.github.com\u002Fcopilot-sdk-workshop\u002F","https:\u002F\u002Fgithub.com\u002Fgithub","https:\u002F\u002Fgithub.com\u002Fabout","How people build software.","2026-11-11 22:34:52",{"source":8,"skillName":5,"commands":35,"prompt":45,"zipAvailable":46},[36,39,42],{"tool":37,"command":38},"npx","npx skills add github\u002Fcopilot-sdk-workshop --skill localizations",{"tool":40,"command":41},"bunx","bunx skills add github\u002Fcopilot-sdk-workshop --skill localizations",{"tool":43,"command":44},"pnpm","pnpm dlx skills add github\u002Fcopilot-sdk-workshop --skill localizations","安装技能，技能地址：https:\u002F\u002Fgithub.com\u002Fgithub\u002Fcopilot-sdk-workshop\u002Fblob\u002Fmain\u002F.github\u002Fskills\u002Flocalizations\u002FSKILL.md",true,[48,54,58,62,66,70,74,78,82,86,92],{"path":49,"resourceType":50,"resourceTypeDesc":51,"sizeBytes":52,"sourceHtmlUrl":53},"plugin.json","REFERENCE","参考文档",185,"https:\u002F\u002Fgithub.com\u002Fgithub\u002Fcopilot-sdk-workshop\u002Fblob\u002Fmain\u002F.github\u002Fskills\u002Flocalizations\u002Fplugin.json",{"path":55,"resourceType":50,"resourceTypeDesc":51,"sizeBytes":56,"sourceHtmlUrl":57},"README.md",1104,"https:\u002F\u002Fgithub.com\u002Fgithub\u002Fcopilot-sdk-workshop\u002Fblob\u002Fmain\u002F.github\u002Fskills\u002Flocalizations\u002FREADME.md",{"path":59,"resourceType":50,"resourceTypeDesc":51,"sizeBytes":60,"sourceHtmlUrl":61},"rules\u002Fde-de.md",13864,"https:\u002F\u002Fgithub.com\u002Fgithub\u002Fcopilot-sdk-workshop\u002Fblob\u002Fmain\u002F.github\u002Fskills\u002Flocalizations\u002Frules\u002Fde-de.md",{"path":63,"resourceType":50,"resourceTypeDesc":51,"sizeBytes":64,"sourceHtmlUrl":65},"rules\u002Fes-es.md",12567,"https:\u002F\u002Fgithub.com\u002Fgithub\u002Fcopilot-sdk-workshop\u002Fblob\u002Fmain\u002F.github\u002Fskills\u002Flocalizations\u002Frules\u002Fes-es.md",{"path":67,"resourceType":50,"resourceTypeDesc":51,"sizeBytes":68,"sourceHtmlUrl":69},"rules\u002Ffr-fr.md",13388,"https:\u002F\u002Fgithub.com\u002Fgithub\u002Fcopilot-sdk-workshop\u002Fblob\u002Fmain\u002F.github\u002Fskills\u002Flocalizations\u002Frules\u002Ffr-fr.md",{"path":71,"resourceType":50,"resourceTypeDesc":51,"sizeBytes":72,"sourceHtmlUrl":73},"rules\u002Fja-jp.md",11622,"https:\u002F\u002Fgithub.com\u002Fgithub\u002Fcopilot-sdk-workshop\u002Fblob\u002Fmain\u002F.github\u002Fskills\u002Flocalizations\u002Frules\u002Fja-jp.md",{"path":75,"resourceType":50,"resourceTypeDesc":51,"sizeBytes":76,"sourceHtmlUrl":77},"rules\u002Fko-kr.md",15159,"https:\u002F\u002Fgithub.com\u002Fgithub\u002Fcopilot-sdk-workshop\u002Fblob\u002Fmain\u002F.github\u002Fskills\u002Flocalizations\u002Frules\u002Fko-kr.md",{"path":79,"resourceType":50,"resourceTypeDesc":51,"sizeBytes":80,"sourceHtmlUrl":81},"rules\u002Fpt-br.md",11359,"https:\u002F\u002Fgithub.com\u002Fgithub\u002Fcopilot-sdk-workshop\u002Fblob\u002Fmain\u002F.github\u002Fskills\u002Flocalizations\u002Frules\u002Fpt-br.md",{"path":83,"resourceType":50,"resourceTypeDesc":51,"sizeBytes":84,"sourceHtmlUrl":85},"rules\u002Fzh-cn.md",11182,"https:\u002F\u002Fgithub.com\u002Fgithub\u002Fcopilot-sdk-workshop\u002Fblob\u002Fmain\u002F.github\u002Fskills\u002Flocalizations\u002Frules\u002Fzh-cn.md",{"path":87,"resourceType":88,"resourceTypeDesc":89,"sizeBytes":90,"sourceHtmlUrl":91},"thumbnail.png","ASSET","资产文件",33267,"https:\u002F\u002Fgithub.com\u002Fgithub\u002Fcopilot-sdk-workshop\u002Fblob\u002Fmain\u002F.github\u002Fskills\u002Flocalizations\u002Fthumbnail.png",{"path":93,"resourceType":88,"resourceTypeDesc":89,"sizeBytes":94,"sourceHtmlUrl":95},"thumbnail.svg",1571,"https:\u002F\u002Fgithub.com\u002Fgithub\u002Fcopilot-sdk-workshop\u002Fblob\u002Fmain\u002F.github\u002Fskills\u002Flocalizations\u002Fthumbnail.svg",[]]