Codex 또는 Claude로 설치 이 Prompt를 복사해 Codex, Claude 또는 다른 어시스턴트에 붙여 넣으면 Skill 페이지를 검토하고 설치를 진행할 수 있습니다.
직접 명령은 검토 Prompt를 거치지 않습니다. 실행하기 전에 소스를 확인하세요.
npx skills add https://github.com/tomevault-io/skills-registry --skill code-review명령은 한 줄로 유지됩니다. 복사하기 전에 가로로 스크롤해 전체 내용을 확인하세요.
로컬 사본을 원하시나요? SkillsMP에서 현재 제공할 수 있는 파일을 다운로드하세요.
| Use when this capability is needed.
> Use when this capability is needed.
Review architecture and API design for the vfs-s3 project. Use when the user mentions @architect, asks to review an issue's design, discuss module boundaries, API shape, or architectural decisions for vfs-s3. Also trigger when the user wants to create an ADR (Architecture Decision Record) or evaluate a technical approach for the project. Intended for dispatch from Codex automation or Claude routines; GitHub trigger phrase: @vfs-s3-bot please prepare design doc Use when this capability is needed.
SOC 직업 분류 기준
SKILL.md 표시 중
| name | code-review |
| description | >- Use when this capability is needed. |
Use this skill to review changed code with a senior-engineer stance: find real bugs and risks introduced by the change, avoid noise, and lead with actionable findings.
Review only the changed behavior unless the user explicitly asks for a broader audit. Do not treat pre-existing issues as findings unless the change makes them newly reachable or worse.
For GitHub PRs, require gh to be installed and authenticated. Use gh for PR
metadata, PR diffs, previous PR comments, and posting comments when requested.
Use git for local diffs, branch history, blame, and full commit SHAs.
Do not run builds, typechecks, linters, or test suites as part of review unless the user asks. CI should catch compiler, formatter, import, and basic lint failures.
Identify the review target:
gh pr view and gh pr diff.Check whether review should proceed:
Gather repository guidance:
AGENTS.md, CLAUDE.md, Cursor rules, contribution docs,
and review guidelines that apply to the modified files.Review through five independent lenses. Use parallel subagents when the runtime supports them; otherwise perform separate passes yourself:
git blame, recent commits, and file history
to understand why nearby code exists.gh.Score each potential issue before reporting it:
0: false positive, pre-existing issue, or unsupported speculation.25: possible issue, but not verified.50: real issue, but minor or unlikely to matter in practice.75: important and likely real, with strong evidence.100: certain, reproducible, and directly caused by the change.For automated PR comments, report only issues scored 80 or higher. In chat,
lead with high-confidence findings and put uncertain observations under open
questions or residual risk.
Do not report:
Follow the host agent's normal response conventions. Do not add vendor-specific signatures, attribution footers, or emoji unless the user asked for that format.
Lead with findings, ordered by severity. For each finding include:
After findings, include open questions or assumptions. Keep summaries brief and
secondary. If no issues are found, say that clearly and mention any review gaps
such as unavailable PR metadata, missing gh authentication, or tests not run.
When posting a GitHub PR comment with gh pr comment, keep it concise and use
full commit SHAs in code links. GitHub links must use this shape:
https://github.com/owner/repo/blob/<full-sha>/path/to/file.ext#L10-L15
Source: Firzus/agent-skills — distributed by TomeVault.