CtrlK
BlogDocsLog inGet started
Tessl Logo

subagent-execution-governance

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。

72

Quality

88%

Does it follow best practices?

Run evals on this skill

Adds up to 20 points to the overall score

View guide

SecuritybySnyk

Passed

No findings from the security scan

The canonical home for this skill is subagent-execution-governance in ooooooooooooooooooop/agent-tools

SKILL.md
Quality
Evals
Security

Quality

Content

77%Weight 40%Scale 1-5

Reviews the quality of instructions and guidance provided to agents. Good implementation is clear, handles edge cases, and produces reliable results.

The body is highly actionable with clear sequenced workflows, validation feedback loops, and concrete templates/budgets, but it is long and treats the audit backstory and plugin specs inline rather than splitting detail into reference files.

Suggestions

Move the P1/P2 plugin specifications into dedicated reference files (e.g. references/subagent-usage-observer.md and references/splice-summarizer.md) and link to them from the body, keeping only a one-line summary inline.

Reference the bundled scripts/workflow_preflight_router.py from the Workflow section (十三) or remove it from the bundle, so no bundle file is orphaned.

Trim the session-audit origin narrative and the reverse-cases table to a brief framing pointer, moving the worked examples to a reference file to reduce token cost on every load.

DimensionReasoningScore

Conciseness

The body is information-dense and does not explain basic concepts Claude already knows, but the 309-line length includes an audit-origin narrative ('本 skill 由一次真实实现会话审计…提炼而成') and a full reverse-cases table that are background color rather than executable protocol and could be tightened.

3 / 5

Actionability

Provides concrete executable artifacts: a contract-snapshot template (GOAL/OWN/MAY READ/IN SCOPE/OUT OF SCOPE/MUST NOT/EXIT/BLOCKED), the three-level read policy with reason codes, the BLOCKED field schema with validity gating, numeric budgets (≤12 reads, ≤6 tool calls, ≤500-word reports), and a 7-item pre-dispatch checklist.

5 / 5

Workflow Clarity

Multi-step processes are explicitly sequenced with validation checkpoints: Discovery→Contract→Implementation→Validation pipeline, the 7-item派发协议清单, the 5-level BLOCKED escalation ladder with feedback loops, and FAIL→repair-contract→re-dispatch recovery with checkpoint persistence.

5 / 5

Progressive Disclosure

Section structure (13 numbered headers) is good and the two audit scripts are referenced with usage, but the P1/P2 plugin specifications are inlined as prose instead of separate reference files, and the bundled scripts/workflow_preflight_router.py is never referenced in the body — content that belongs in separate files is inline and a bundle file is orphaned.

3 / 5

Total

16

/

20

Passed

Description

100%Weight 40%Scale 1-5

Based on the skill's description, can an agent find and select it at the right time? Clear, specific descriptions lead to better discovery.

The description is third-person, concrete, and explicitly pairs what-it-does with when-to-use-it and when-not-to-use-it, with natural trigger phrases. It is among the strongest patterns in the reference examples.

DimensionReasoningScore

Specificity

Lists multiple concrete mechanisms — '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' — comprehensive and specific rather than abstract.

5 / 5

Completeness

Explicitly answers both 'what' (契约型治理 upgrade with the listed mechanisms) and 'when' ('用于…' concrete triggers plus '不用于…' exclusions), matching the anchor that requires both with concrete trigger phrases.

5 / 5

Trigger Term Quality

Natural phrases users would actually say are covered with synonyms: '派发实现型子代理(写代码/写文件类任务)', '审计子代理空转(反复探查不产出、并行写冲突、上下文膨胀)', '复盘子代理烧了大量 token 却没产出的会话'.

5 / 5

Distinctiveness Conflict Risk

Clear niche (subagent execution governance / contract-based dispatch) with distinct triggers; the body even delineates boundaries against adjacent skills (execution-discipline, decision-gates), minimizing wrong-skill activation.

5 / 5

Total

20

/

20

Passed

Validation

93%

Checks the skill against the spec for correct structure and formatting. All validation checks must pass before discovery and implementation can be scored.

Validation15 / 16 Passed

Validation for skill structure

CriteriaDescriptionResult

frontmatter_unknown_keys

Unknown frontmatter key(s) found; consider removing or moving to metadata

Warning

Total

15

/

16

Passed

Repository
ooooooooooooooooooop/personal-ai
Reviewed

Table of Contents

Is this your skill?

If you maintain this skill, you can claim it as your own. Once claimed, you can manage eval scenarios, bundle related skills, attach documentation or rules, and ensure cross-agent compatibility.