[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"skill-48c1d500-33b2-4b43-9cb8-c87abeddbf0f":3,"$fXegpFWo1F0hs_JhX3Vr2I6M4OfiZcBrG8wXiJFX8X4U":43},{"id":4,"title":5,"description":6,"categoryId":7,"moduleId":8,"tags":9,"prompt":10,"icon":11,"source":12,"sourceUrl":13,"authorId":14,"authorName":15,"isPublic":16,"stars":17,"runs":18,"createdAt":19,"updatedAt":19,"module":20,"category":27,"packages":34},"48c1d500-33b2-4b43-9cb8-c87abeddbf0f","skill-writer","根据《代理技能规范》创建和改进代理技能。在需要创建、编写或更新技能时使用。","cat_life_career","mod_other","sickn33,other","---\nname: skill-writer\ndescription: Create and improve agent skills following the Agent Skills specification. Use when asked to create, write, or update skills.\nrisk: unknown\nsource: community\n---\n\n# Skill Writer\n\nUse this as the single canonical workflow for skill creation and improvement.\nPrimary success condition: maximize high-value input coverage before authoring so the resulting skill has minimal blind spots.\n\nLoad only the path(s) required for the task:\n\n| Task | Read |\n|------|------|\n| Set skill class and required dimensions | `references\u002Fmode-selection.md` |\n| Apply writing constraints for depth vs concision | `references\u002Fdesign-principles.md` |\n| Select structure pattern for this skill | `references\u002Fskill-patterns.md` |\n| Select workflow orchestration pattern for process-heavy skills | `references\u002Fworkflow-patterns.md` |\n| Select output format pattern for deterministic quality | `references\u002Foutput-patterns.md` |\n| Choose workflow path and required outputs | `references\u002Fmode-selection.md` |\n| Load representative synthesis examples by skill type | `references\u002Fexamples\u002F*.md` |\n| Synthesize external\u002Flocal sources with depth gates | `references\u002Fsynthesis-path.md` |\n| Author or update SKILL.md and supporting files | `references\u002Fauthoring-path.md` |\n| Optimize skill description and trigger precision | `references\u002Fdescription-optimization.md` |\n| Iterate using positive\u002Fnegative\u002Ffix examples | `references\u002Fiteration-path.md` |\n| Evaluate behavior and compare baseline vs with-skill (opt-in quantitative) | `references\u002Fevaluation-path.md` |\n| Register and validate skill changes | `references\u002Fregistration-validation.md` |\n\n## Step 1: Resolve target and path\n\n1. Resolve target skill path and intended operation (`create`, `update`, `synthesize`, `iterate`).\n2. Read `references\u002Fmode-selection.md` and select the required path(s).\n3. Classify the skill (`workflow-process`, `integration-documentation`, `security-review`, `skill-authoring`, `generic`).\n4. Ask one direct question if class or depth requirements are ambiguous; otherwise state explicit assumptions.\n\n## Step 2: Run synthesis when needed\n\nRead `references\u002Fsynthesis-path.md`.\n\n1. Collect and score relevant sources with provenance.\n2. Apply trust and safety rules when ingesting external content.\n3. Produce source-backed decisions and coverage\u002Fgap status.\n4. Load one or more profiles from `references\u002Fexamples\u002F*.md` when the skill is hybrid.\n5. Enforce baseline source pack for skill-authoring workflows.\n6. Enforce depth gates before moving to authoring.\n\n## Step 3: Run iteration first when improving from outcomes\u002Fexamples\n\nRead `references\u002Fiteration-path.md` first when selected path includes `iteration` (for example operation `iterate`).\n\n1. Capture and anonymize examples with provenance.\n2. Re-evaluate skill behavior against working and holdout slices.\n3. Propose improvements from positive\u002Fnegative\u002Ffix evidence.\n4. Carry concrete behavior deltas into authoring.\n\nSkip this step when selected path does not include `iteration`.\n\n## Step 4: Author or update skill artifacts\n\nRead `references\u002Fauthoring-path.md`.\n\n1. Write or update `SKILL.md` in imperative voice with trigger-rich description.\n2. Create focused reference files and scripts only when justified.\n3. Follow `references\u002Fskill-patterns.md`, `references\u002Fworkflow-patterns.md`, and\n   `references\u002Foutput-patterns.md` for structure and output determinism.\n4. For authoring\u002Fgenerator skills, include transformed examples in references:\n   - happy-path\n   - secure\u002Frobust variant\n   - anti-pattern + corrected version\n\n## Step 5: Optimize description quality\n\nRead `references\u002Fdescription-optimization.md`.\n\n1. Validate should-trigger and should-not-trigger query sets.\n2. Reduce false positives and false negatives with targeted description edits.\n3. Keep trigger language generic across Codex and Claude.\n\n## Step 6: Evaluate outcomes\n\nRead `references\u002Fevaluation-path.md`.\n\n1. Run a lightweight qualitative check by default (recommended).\n2. For integration\u002Fdocumentation and skill-authoring skills, include the concise depth rubric from `references\u002Fevaluation-path.md`.\n3. Run deeper eval playbook and quantitative baseline-vs-with-skill only when requested or risk warrants it.\n4. Record outcomes and unresolved risks.\n\n## Step 7: Register and validate\n\nRead `references\u002Fregistration-validation.md`.\n\n1. Apply repository registration steps.\n2. Run quick validation with strict depth gates.\n3. Reject shallow outputs that fail depth gates or required artifact checks.\n\n## Output format\n\nReturn:\n\n1. `Summary`\n2. `Changes Made`\n3. `Validation Results`\n4. `Open Gaps`\n\n## When to Use\nUse this skill when tackling tasks related to its primary domain or functionality as described above.\n\n## Limitations\n- Use this skill only when the task clearly matches the scope described above.\n- Do not treat the output as a substitute for environment-specific validation, testing, or expert review.\n- Stop and ask for clarification if required inputs, permissions, safety boundaries, or success criteria are missing.\n","","imported","https:\u002F\u002Fgithub.com\u002Fsickn33\u002Fantigravity-awesome-skills","user_system_seed","SkillOPIC",true,159,1703,"2026-05-16 13:41:01",{"id":8,"name":21,"slug":22,"icon":23,"description":24,"sort":25,"createdAt":26},"其他","other","mdi-page-next-outline","其他类型Skill",5,"2026-05-16 12:53:40",{"id":7,"name":28,"slug":29,"icon":30,"description":31,"moduleId":8,"sort":32,"skillCount":33,"createdAt":26},"职场发展","career","mdi-briefcase-outline","面试准备、简历优化、职业规划",4,575,[35],{"id":36,"skillId":4,"version":37,"fileName":38,"fileSize":39,"filePath":40,"fileHash":41,"manifest":42,"createdAt":19},"e974d9e4-8d22-4872-8702-2bb0b9e40b79","1.0.0","skill-writer.zip",9891,"uploads\u002Fskills\u002F48c1d500-33b2-4b43-9cb8-c87abeddbf0f\u002Fskill-writer.zip","90ed841506dce8c1beb64c59f487cca9f506f98bbfe9e2bf8a7c3880133ec7e1","[{\"path\":\"SKILL.md\",\"isDirectory\":false,\"size\":5063},{\"path\":\"references\u002Fauthoring-path.md\",\"isDirectory\":false,\"size\":872},{\"path\":\"references\u002Fdescription-optimization.md\",\"isDirectory\":false,\"size\":973},{\"path\":\"references\u002Fdesign-principles.md\",\"isDirectory\":false,\"size\":1165},{\"path\":\"references\u002Fevaluation-path.md\",\"isDirectory\":false,\"size\":741},{\"path\":\"references\u002Fexamples\u002Fworkflow-process.md\",\"isDirectory\":false,\"size\":638},{\"path\":\"references\u002Fiteration-path.md\",\"isDirectory\":false,\"size\":875},{\"path\":\"references\u002Fmode-selection.md\",\"isDirectory\":false,\"size\":1650},{\"path\":\"references\u002Foutput-patterns.md\",\"isDirectory\":false,\"size\":651},{\"path\":\"references\u002Fregistration-validation.md\",\"isDirectory\":false,\"size\":932},{\"path\":\"references\u002Fskill-patterns.md\",\"isDirectory\":false,\"size\":969},{\"path\":\"references\u002Fsynthesis-path.md\",\"isDirectory\":false,\"size\":872},{\"path\":\"references\u002Fworkflow-patterns.md\",\"isDirectory\":false,\"size\":964}]",{"code":44,"message":45,"data":46},200,"success",{"items":47,"stats":48,"page":51},[],{"averageRating":49,"totalRatings":49,"ratingCounts":50},0,[49,49,49,49,49],{"limit":52,"offset":49,"hasMore":53,"nextOffset":52,"ratedOnly":16},15,false]