retro

仓库创建 2026年7月5日最近提交 22 天前SkillHot 收录 22 天前
▸ 精选理由

便于从对话中形成可检索的经验库,支持持续改进。

▸ 风险提示

会写入经验卡/文档,注意不要泄露敏感对话内容。

这个 Skill 做什么

会话复盘,挖掘修正与审查结论并整理为去重的经验卡片。

把一次会话里用户的改动、审查结论和排错根因都挖出来,整理成去重的经验卡片,方便以后查阅和复用。适合在会议、代码评审或长会话结束后做复盘,把零散结论结构化成可搜索的教训条目。会自动合并重复项并清理过期信息,保持经验库干净、不膨胀。

▸ 展开 SKILL.md 英文原文

Session post-mortem: mine the conversation for user corrections, confirmed review findings, and debug root causes; distill them into lesson cards; merge duplicates and prune stale ones.

开发编程回顾知识管理课程卡片通用
0
Stars
0
Forks
15
仓库内 Skill
+0
7 日增星
安装 / 使用
给你的 Agent 一句话(通用)
帮我安装这个 skill:https://raw.githubusercontent.com/eduardkumskyi/octo/master/skills/retro/SKILL.md
或 curl 直取 SKILL.md
curl -fsSL "https://raw.githubusercontent.com/eduardkumskyi/octo/master/skills/retro/SKILL.md"
SKILL.MD 节选查看完整文件 ↗
## Progress Contract

Register these steps as a native task list at Step 1, before beginning.
Report progress as "N steps remaining, size class S/M/L" — never wall-clock ETAs.

Steps: (1) mine-session, (2) write-cards, (3) curate.

## Workflow

### Step 1 — Mine session

Dispatch **three subagents in one message** — one per signal type — to mine the conversation
history concurrently. Sequential mining of independent signal types is a defect:

- **(a) User corrections** — any place the user rejected an output, flagged a mistake, or
  corrected an assumption. Extract the erroneous pattern and the correct approach.
- **(b) Confirmed review findings** — findings that survived adversarial verific
via SKILL·HOT · 数据来自 GitHub 公开信息 · 原文版权归作者所有