Menu
"PDD熵减智能体，持续监控和偿还技术债务，防止系统腐化。触发：熵减、技术债务、代码清理、文档更新、架构对齐、垃圾回收、entropy reduction。"
The review-first prompt is selected by default. You can switch to a direct command or download a local copy.
Read SKILL.md and any companion files shown by SkillsMP before deciding whether to install.
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.
A direct command skips the review prompt. Inspect the source before running it.
npx skills add https://github.com/wonderslife/pdd-skills-v3 --skill pdd-entropy-reductionThe command stays on one line. Scroll horizontally to inspect it before copying.
Prefer a local copy? Download the files currently available to SkillsMP.
Based on SOC occupation classification
Showing SKILL.md
| name | pdd-entropy-reduction |
| description | "PDD熵减智能体，持续监控和偿还技术债务，防止系统腐化。触发：熵减、技术债务、代码清理、文档更新、架构对齐、垃圾回收、entropy reduction。" |
"Technical debt is like a high-interest loan: better to repay in small amounts than let it accumulate." —— OpenAI Harness Engineering
目标：通过定期运行智能体发现文档不一致或架构违反，对抗系统熵增与衰减。
熵增 vs 熵减：代码腐化(重复/过长/命名)→重构；文档过时→同步；债务累积(TODO/临时方案)→偿还；架构漂移(依赖方向/边界)→对齐；测试不足→补充。
协调器(pdd-entropy-reduction)负责调度和聚合结果，不直接执行具体操作：
触发(手动/定时/事件) → 扫描(文档/代码/架构) → 分析(熵评分/分类/优先级) → 执行(自动修复/创建PR/更新文档)
熵评分系统(0-100，100=最有秩序)：90-100优秀维持 | 70-89良好小幅改进 | 50-69一般计划清理 | 30-49警告优先处理 | 0-29危急紧急重构
references/entropy-report-template.md，含熵评分/问题列表(按优先级)/修复建议/预估工作量entropy-config.yaml）entropy_reduction:
triggers: { schedule: "0 2 * * *", on_commit: false, on_pr_merge: true }
detection:
docs: { enabled: true, paths: ["docs/", "*.md"], max_age_days: 30 }
architecture: { enabled: true, layers: ["types","config","repo","service","runtime","ui"] }
code: { enabled: true, max_file_lines: 300, max_function_lines: 50 }
execution: { auto_fix: true, create_pr: true, max_pr_per_run: 5 }
每次执行后生成报告存到 docs/entropy-reports/：# Entropy Reduction Report - YYYY-MM-DD + 熵评分 + 问题列表(Critical/Warning/Info) + 修复建议与执行结果。
PDD正向流程(PRD→功能提取→规格→实现→验收) ←→ PDD熵减流程(检测→报告→执行→自动修复→PR创建)