一键导入
address-review
Evaluate PR review comments for validity and address confirmed ones
用 Codex 或 Claude 帮你安装 复制这段 Prompt,粘贴到 Codex、Claude 或其他助手里,让它检查 Skill 页面并帮你完成安装。
菜单
Evaluate PR review comments for validity and address confirmed ones
用 Codex 或 Claude 帮你安装 复制这段 Prompt,粘贴到 Codex、Claude 或其他助手里,让它检查 Skill 页面并帮你完成安装。
基于 SOC 职业分类
| name | address-review |
| description | Evaluate PR review comments for validity and address confirmed ones |
You are helping a developer respond to review comments on a pull request in the embedded-services repository.
Do NOT assume review comments are correct. Reviewers (both human and AI) can be wrong. Independently evaluate each comment against the actual code and context before proposing changes.
Read every review comment and the surrounding code. Present your analysis in a table:
| # | Comment | Location | Valid? | Reasoning | Proposed Fix |
|---|---|---|---|---|---|
| 1 | Reviewer's comment (summarized) | src/file.rs:42 ([link]) | ✅ Yes / ❌ No / ⚠️ Partial | Why you agree or disagree, with code snippet | What you would change (or "No change needed") |
Include a short code snippet in the Reasoning or Proposed Fix column to show the relevant context:
// current code
let val = buf[i];
// proposed fix
let val = buf.get(i).ok_or(MyError::InvalidParams)?;
After presenting the table, ask the developer which comments to address by number. Do not make any changes until confirmed.
Implement only the confirmed fixes. For comments marked invalid that the developer still wants addressed, discuss the tradeoff before proceeding.
no_std context, strict clippy policy, and feature-gated code when evaluating suggestions.unwrap(), panic!(), or direct indexing, flag it as invalid — these are denied by clippy config.Orchestrates cargo-vet supply chain audits for Rust crates. Use this skill when asked to audit dependencies, review supply chain security, certify crates with cargo vet, or assess the trustworthiness of imported audit sources.
AI-assisted code review that produces a structured summary for human reviewers