Menu
Distills iteration lessons, categorizes them, and injects them into the most relevant Skill.md files with deduplication.
Install with Codex or Claude Copy this prompt, paste it into Codex, Claude, or another assistant, and let it review the skill page and install it for you.
Based on SOC occupation classification
MiniHES 后端开发规范。当用户进行 FastAPI 后端开发、API 设计、数据库操作、 DLMS 协议实现、数据采集任务开发时自动激活。 涵盖分层架构、编码原则、安全要求、数据库规范。
MiniHES 代码审查。当用户请求审查代码、review PR、检查变更时激活。 支持本地变更审查和指定文件审查，输出结构化审查报告。 引用 backend-checklist.md 和 frontend-checklist.md 作为审查依据。
修复代码格式和 lint 问题。当用户在提交前需要修复代码风格、 排查 lint 错误、或说 "fix"、"格式化"、"lint" 时激活。
MiniHES 前端开发规范。当用户进行 Vue 3 前端开发、组件设计、 页面布局、样式调整时自动激活。 涵盖技术栈、组件设计原则、设计规范、Ant Design Vue 用法。
Systematic code review patterns covering security, performance, maintainability, correctness, and testing
创建 Pull Request。当用户说 "创建 PR"、"提 PR"、"提交合并请求" 时激活。 确保遵循 conventional commit、模板规范、安全护栏。
| name | Knowledge Distillation |
| description | Distills iteration lessons, categorizes them, and injects them into the most relevant Skill.md files with deduplication. |
| id | knowledge_distillation |
| dimensions | ["classification","deduplication","skill_injection"] |
| tools | ["markdown","file_read","file_write","code_search"] |
| constraints | ["must_identify_target_skill_accurately","must_check_existing_lessons_to_prevent_duplicates","must_maintain_consistent_lesson_format","must_only_modify_lessons_learned_section"] |
| input_schema | {"type":"object","required":["task_context","execution_artifacts","available_skills"],"properties":{"task_context":{"type":"object","description":"Original goals, requirements, and user feedback."},"execution_artifacts":{"type":"object","description":"Logs, diffs, errors, and retry traces."},"available_skills":{"type":"array","items":{"type":"string"},"description":"List of available Skill IDs or paths that can be updated."}}} |
| output_schema | {"type":"object","required":["updates"],"properties":{"updates":{"type":"array","items":{"type":"object","required":["target_skill_id","action","lesson_content"],"properties":{"target_skill_id":{"type":"string"},"action":{"type":"string","enum":["add","merge","update","ignore"]},"lesson_content":{"type":"object","required":["context","lesson","impact","evidence"],"properties":{"context":{"type":"string"},"lesson":{"type":"string"},"impact":{"type":"string"},"evidence":{"type":"string"}}}}}}}} |
| skill_type | cognitive |
| side_effects | ["file_write"] |
| deterministic | false |
| testable | true |
| metadata | {"execution_mode":"analysis","execution_tools":["file_read","file_write","code_search"],"execution_capabilities":["skill_categorization","knowledge_deduplication","automated_skill_evolution"]} |
Analyze iteration data to extract reusable lessons, categorize them into the most relevant skills, and perform idempotent updates to Skill.md files.
Determine which specific skill (e.g., python_engineering, schema_design) a lesson most accurately belongs to.
Analyze existing "Lessons Learned" sections in target skills to ensure new entries provide unique value and don't replicate existing knowledge.
Precisely inject or merge structured lessons into the target Skill.md files while preserving the overall file structure.
Identifies obvious lessons and adds them to a generic "General" skill or log.
Correctly categorizes lessons into specific functional skills and avoids exact duplicates.
Performs semantic deduplication, merges related lessons into more powerful generalized rules, and maintains a highly optimized Skill library.
