AGENT SKILL
doctrine
The operating doctrine, self-contained in this folder. Invoke BEFORE - delegating work to a subagent via the Agent tool (picking model tier, writing the dispatch prompt) or handling a subagent that failed (escalate/de-escalate); deciding whether to retry, escalate, switch approach, or ask the user; reporting a nontrivial task complete (the done-gate); editing this doctrine or CLAUDE.md; recording a lesson about the harness.
使用场景Skill与智能体开发
- License 许可
- MIT
- 最近更新
- 2026-07-19
- Stars
- ★ 244
- Forks
- 35
01 前言 Front Matter
- name
- doctrine
- description
- The operating doctrine, self-contained in this folder. Invoke BEFORE - delegating work to a subagent via the Agent tool (picking model tier, writing the dispatch prompt) or handling a subagent that failed (escalate/de-escalate); deciding whether to retry, escalate, switch approach, or ask the user; reporting a nontrivial task complete (the done-gate); editing this doctrine or CLAUDE.md; recording a lesson about the harness.
- user-invocable
- false
02 资源文件 14
- 参考文档00-diagnosis.md7.1 KB
- 参考文档01-model-dispatch.md9.7 KB
- 参考文档02-judgment.md8.8 KB
- 参考文档03-delegation-templates.md6.2 KB
- 参考文档04-maintenance.md6.6 KB
- 参考文档05-letter.md6.2 KB
- 参考文档06-lessons.md5.2 KB
- 参考文档INSTALL.md2.7 KB
- 许可证LICENSE.txt11.1 KB
- 参考文档references/output-patterns.md1.8 KB
- 参考文档references/workflows.md818 B
- 脚本scripts/init_skill.py10.6 KB
- 脚本scripts/package_skill.py3.2 KB
- 脚本scripts/quick_validate.py3.4 KB
03 SKILL.md 正文
Doctrine — the operating institution, one folder
All referenced files live in this skill's folder; read them with relative paths from here.
The three always-on laws live in the user's ~/.claude/CLAUDE.md router; this bundle carries
the depth. Content lives in the numbered files — this SKILL.md only routes; never duplicate
rule text into it.
Route by moment
| Moment | Read (in this folder) |
|---|---|
| About to delegate / pick a model tier / a subagent failed | 01-model-dispatch.md, then fill a template from 03-delegation-templates.md |
| Retry? escalate? done? ask the user? direction feels wrong? | 02-judgment.md — five rubrics, each with a right and wrong example |
| About to claim a nontrivial task is done | 02-judgment.md Rubric 2 (the done-gate) — every box, before you say "done" |
Editing any file here or CLAUDE.md; recording a lesson |
04-maintenance.md — preserve history first (git commit or backup), self-edit boundaries, formats |
| Confused by the harness itself | 06-lessons.md first (it may already be paid for), then 00-diagnosis.md |
| First session in a new environment, or sharing this bundle | INSTALL.md |
| Why these rules exist; environment-specific warnings | 00-diagnosis.md, 05-letter.md |
Dispatch pre-send checklist (all four, or don't send)
- [ ] Goal + motive, acceptance criteria, report format all present (the dispatch triple,
01). - [ ]
modelchosen explicitly — mechanical→haiku, default→sonnet, hard→opus— not defaulted. - [ ] Absolute paths and exact symbols named — the subagent has zero of your context.
- [ ] Long outputs directed to a file, path returned — never pasted back into the conversation.
Quick reference
Would raw output ≫ the answer? → delegate (Explore for read-only searches)
Every dispatch: goal+motive / accept / report-format
Model: mechanical→haiku default→sonnet hard→opus hardest→fable (if available — see 01)
Fail: small errs once→up mid fails 2×→opus+trace solved→down 2 escalation rounds max
Report: conclusion first · file:line · big artifact→file+path · blocked→say why
Verify: fresh context · read-back files · run code · refute high-stakes
