Structured code review through the shared OpenClaw agent-skills installation.
中文简介正在准备中。当前显示原文说明。
Clean only the current Enterprise diff before autoreview, preserving behavior, security boundaries, required TODOs, and useful test-intent comments.
中文简介正在准备中。当前显示原文说明。
查看安装方法安装前可以查看交给 AI 代理的指令内容。
Keep cleanup confined to the current diff and strictly behavior-neutral. Use it before requested autoreview; cleanup does not replace correctness or safety review.
For readability choices in the current diff, consult the relevant sections of Readable code on functions, values, composition, and state ownership. Apply those examples within this skill's behavior-neutral scope.
Establish the intended branch base from the PR or repository metadata. Inspect committed changes from its merge base to HEAD, plus staged, unstaged, and relevant untracked changes. Preserve unrelated work; never sweep the repository or change the base to make the cleanup larger. Read surrounding code to judge each hunk.
Look for:
as any or as unknown as T that hide a type problem.These patterns require judgment. Apply only trivial, demonstrably behavior-neutral edits. Removing a guard, catch, retry, fallback, or type cast can change behavior or expose a real type-contract defect: leave uncertain or functional changes in place and report them for separate review. Do not hide errors with wider types or replace security checks with assumptions.
Keep real fail-closed checks, authorization and credential boundaries, configuration rejection, and attributable audit behavior. Redundancy alone is not evidence that a security check is removable.
Preserve deferred-work TODOs while the capability remains unimplemented. Preserve integration-test comments that explain scenarios, non-obvious setup, expected outcomes, or security invariants; remove only narration that contributes no such information. Do not label permanent security boundaries as temporary work.
Use enterprise-testing for proportional
verification of edits and run git diff --check. If behavior neutrality cannot
be established, report the candidate instead of applying it as style work.
Summarize in one to three sentences what changed and any non-trivial candidates
left for the author.
Adapted from OpenClaw; see source and intentional adaptations.
还没有评价。欢迎分享你的使用感受。
无需登录即可发表。欢迎分享使用感受、提问,或说明在哪些环境下无法运行。
还没有评论。来写第一条吧。
Structured code review through the shared OpenClaw agent-skills installation.
中文简介正在准备中。当前显示原文说明。
Assess designs, public interfaces, state ownership, readability, and refactor proposals when the task calls for design review or structural simplification; not for unrelated routine edits.
中文简介正在准备中。当前显示原文说明。
Select proportional Enterprise validation and diagnose exact CI runs, distinguishing product, harness, infrastructure, and credential failures.
中文简介正在准备中。当前显示原文说明。
Develop OpenClaw Enterprise changes with proportional verification and source-backed flow documentation for non-trivial behavior.
中文简介正在准备中。当前显示原文说明。
Create or refine readable Mermaid diagrams for changes, architecture, lifecycles, and dependencies in Markdown documents and PR descriptions.
中文简介正在准备中。当前显示原文说明。
Run named OpenClaw Enterprise end-to-end integration scenarios against real supported installations. Use only when explicitly invoked.
中文简介正在准备中。当前显示原文说明。