AGENT SKILL

ai-review

Run an independent adversarial correctness review against a specification

License 许可
MIT
最近更新
2026-07-29
Stars
★ 0
Forks
0

01 前言 Front Matter

name
ai-review
description
Run an independent adversarial correctness review against a specification
argument-hint
[spec-path]
disable-model-invocation
true

03 SKILL.md 正文

Delegate to the correctness-reviewer agent: review the current implementation against $ARGUMENTS. Determine the base reference from the current run manifest or AI_BASE_REF.

For risk R2 and above, also delegate to security-reviewer in a separate context and preserve both sets of findings verbatim.

04 同仓库更多 Skills