subagent-execution-governance SKILL.md skillA
subagent-execution-governance from ooooooooooooooooooop/agent-tools is an agent skill: a SKILL.md folder an AI agent loads when a task calls for it.
mdr add ooooooooooooooooooop/agent-tools/subagent-execution-governance@v0.2.0Pins v0.2.0 in mdr.lock. Pin the fingerprint instead.
- Audit grade
- A, 17 of 17 checks passed
- Kind
- SKILL.md skill
- Repository
- ooooooooooooooooooop/agent-tools
- Path
skills/subagent-execution-governance/SKILL.md- Versions recorded
- 5
- Latest version
- v0.2.0, committed Sep 1, 2026
- First version recorded
- Aug 23, 2026
- Size
- 19,915 bytes
- Repository stars
- 4
- License
- MIT
把子代理派发从"监督型治理"升级为"契约型治理"：探查/实现/验证角色分离、契约快照（含负空间）、三级有界读取、单写者文件所有权、行动门槛替代轮次门槛、结构化 BLOCKED、token/工具/轮次预算、结构化结果回收与 repair-contract 恢复。用于派发实现型子代理（写代码/写文件类任务）、设计"谁读什么/谁写什么/何时算阻塞"的执行协议、审计子代理空转（反复探查不产出、并行写冲突、上下文膨胀）、或复盘"子代理烧了大量 token 却没产出"的会话。不用于纯问答、单轮小修改、或已有成熟任务流程的普通委派。.
Collected from public GitHub. Every version is saved exactly as it was and never changes. Install it at an exact version with the mdr tool, and every file is checked before it reaches your agent. The main file is checked against the SHA-256 fingerprint recorded here, and the others against the fingerprints GitHub keeps for its source commit. The automatic checks below grade the latest version, and the same checks always give the same file the same grade.
What the file says
# 子代理执行治理（Subagent Execution Governance） > 本 skill 由一次真实实现会话审计（session-33b69ec9-1abf-454d-9555-f2d4031a9453）提炼而成。那次会话 3 小时、17 个子代理、执行层消耗 309 万输入 token（主会话 4 倍），`b21b9211` 一个子代理读了 65 个文件后仍被催"直接实现"，主会话被迫 28 次 `send_message` 催促 + 8 次 `interrupt_agent`，最终目标未 complete、会话中途悬死。**根因不是"监督不够凶"，而是子代理的工作边界、信息边界、写权限和退出条件没有工程化。** ## 核心原则 > 不要从"自由探索"直接走向"实现代理完全失去自主权"；正确形态是 **契约驱动 + 有界自治 + 结构化升级**。 > > 目标不是"主代理不断管子代理"，而是 **系统规则让正常子代理根本不需要被管**。 失败因果链（本 skill 要打断的）： ```text 任务设计错误（探查/实现不分） → 执行边界过宽（无限读取授权） → 监督不得不频繁介入（催促/中断/重派） → send_message / interrupt 注入上下文 → 上下文继续膨胀 → 缓存命中率恶化 → token 爆炸 ``` ## 一、角色分离（Discovery → Contract → Implementation → Validation） 禁止让一个子代理同时承担"先研究、再实现、再测试"四种认知模式： | 角色 | 读 | 写 | 产出 | |------|----|----|------| | **Discovery**（探查） | 自由只读 | 禁止 | 契约快照草案（≤15 条 file:line 事实） | | **Contract Builder**（主会话） | 验证契约 | 写契约文件 | 最终契约快照（含负空间） | | **Implementation**（实现） | 有界读取（三级） | 只写自有文件 | 代码 + 变更清单 + 状态 | | **Validation**（验证） | 只读目标文件/测试 | 禁止（或只改测试） | 测试结果 + 差异报告 | **主会话承担 Contract Builder 与 Orchestrator**：探查结果必须经主会话验证、确认文件无冲突、写入契约快照后，才能派实现子代理。**禁止**探查子代理直接产出实现任务。 ## 二、契约快照（Implementation Contract） 每个实现子代理派发前必须有一份契约快照，包含**正空间和负空间**： ```text GOAL: 一句话目标（可验收） …
Read the whole file at its exact version.
How to install
mdr add ooooooooooooooooooop/agent-tools/subagent-execution-governance@sha256:db3ba9c54c446be2The line at the top of this page pins the label v0.2.0, which is what most people want. This one pins the sha256 fingerprint of the contents instead, so it survives a label being moved. Either way the exact version is written to mdr.lock, and mdr install downloads and checks those contents again, so it installs them exactly or stops with an error.
[](https://markdownregistry.com/a/art_eq6fhytjoqqdryxu)
3 badge image fetches in 30 days. What the badge shows.
Versions
The registry has recorded 5 versions of this file, between Aug 23, 2026 and Sep 1, 2026, about one every 2 days.
| version | committed | commit | size | audit | |
|---|---|---|---|---|---|
| v0.2.0 latest | 2026-09-01 | 75025e2 | 19,915 B | A | view · diff |
| v0.2.0 | 2026-08-28 | 278776e | 17,309 B | A | view · diff |
| v0.1.0 | 2026-08-25 | 1dcadfc | 15,856 B | A | view · diff |
| v0.1.0 | 2026-08-23 | fc0b525 | 14,851 B | A | view · diff |
| v0.1.0 | 2026-08-23 | 8753690 | 14,852 B | A | view |
Audit of the latest version
- pass: Frontmatter block present
- pass: Frontmatter declares a name
- pass: Frontmatter declares a description
- pass: Size between 200 bytes and 200 KB (19915 bytes)
- pass: No zero-width or bidi control characters
- pass: No instruction hidden inside an HTML comment
- pass: No link to an exfiltration or paste host
- pass: No credential-shaped string
- pass: No instruction to send local credentials anywhere
- pass: No text hidden with inline styles
- pass: No prompt-injection phrasing
- pass: No curl or wget piped into a shell
- pass: No recursive delete of root, home or parent
- pass: No instruction to read or print local credentials
- pass: No base64 blob over 200 characters
- pass: No link to a raw IP address
- pass: No script tag
Source
ooooooooooooooooooop/agent-tools · 4 stars · license MIT · pushed 2026-10-23 · branch main
API
GET https://markdownregistry.com/api/v1/artifacts/art_eq6fhytjoqqdryxu GET https://markdownregistry.com/api/v1/resolve?ref=ooooooooooooooooooop/agent-tools/subagent-execution-governance GET https://markdownregistry.com/api/v1/blob/db3ba9c54c446be2985d64654bc5f388aaddac48e2734f6ccc4e9b8296735bf9
Your AI assistant does the legwork. You hear about the deals worth your word. Hand yours the standing instructions at modelranch.com and it joins the network that reads files like this one.
Guides: the SKILL.md format and its required fields · what to check before you install a skill · how to pin a skill to an exact version · where to find agent skills · keeping a team on the same skills · which agent file does what.
More from ooooooooooooooooooop/agent-tools
Every file in ooooooooooooooooooop/agent-tools