code-review
Language-agnostic code review skill. Reads project conventions from CLAUDE.md and applies parameterized quality checks.
Codex 또는 Claude로 설치 이 Prompt를 복사해 Codex, Claude 또는 다른 어시스턴트에 붙여 넣으면 Skill 페이지를 검토하고 설치를 진행할 수 있습니다.
메뉴
Language-agnostic code review skill. Reads project conventions from CLAUDE.md and applies parameterized quality checks.
Codex 또는 Claude로 설치 이 Prompt를 복사해 Codex, Claude 또는 다른 어시스턴트에 붙여 넣으면 Skill 페이지를 검토하고 설치를 진행할 수 있습니다.
SOC 직업 분류 기준
Manage project board — issues, sprints, status tracking, acceptance verification, and release management. Supports GitHub Projects, Jira, and YouTrack via pluggable providers.
Manage project board — issues, sprints, status tracking, acceptance verification, and release management. Supports GitHub Projects, Jira, and YouTrack via pluggable providers.
Happy path smoke testing — navigates all configured endpoints, checks HTTP status and server logs for errors, and optionally auto-creates GitHub issues for failures. Config-driven via CC_SMOKE_TEST_* variables.
Manage GitHub Project board — issues, sprints, status tracking, acceptance verification, and release management. White-labeled template for any project.
Check for and apply cognitive-core framework updates. Compares installed agents, skills, and hooks against the framework source, shows available updates, and safely applies them.
Check for and apply cognitive-core framework updates. Compares installed agents, skills, and hooks against the framework source, shows available updates, and safely applies them.
| name | code-review |
| description | Language-agnostic code review skill. Reads project conventions from CLAUDE.md and applies parameterized quality checks. |
| user-invocable | true |
| allowed-tools | Read, Grep, Glob |
| catalog_description | Language-agnostic code review with project-aware convention checks. |
Provides structured code review by reading conventions from your project's
CLAUDE.md and cognitive-core.conf. All checks are parameterized.
$ARGUMENTS -- file or directory to reviewCLAUDE.md at the project root for coding standardscognitive-core.conf for language and architecture settings:
CC_LANGUAGE -- primary languageCC_ARCHITECTURE -- architecture pattern (ddd, mvc, clean, etc.)CC_LINT_COMMAND -- configured lint commandDetermine the architectural layer of the file being reviewed based on its path
and the configured CC_ARCHITECTURE pattern. Common layers:
| Check | Severity |
|---|---|
| File follows project naming conventions | ERROR |
| Error handling present (not bare catch/rescue/except) | ERROR |
| No hardcoded secrets or credentials | ERROR |
| Functions/methods have clear single responsibility | WARN |
| Magic numbers or strings extracted to constants | WARN |
| Input validation on public interfaces | WARN |
| Check | Severity |
|---|---|
| Layer dependencies follow configured pattern | ERROR |
| No infrastructure code in domain/model layer | ERROR |
| Controllers delegate to services (thin controllers) | WARN |
| Data access goes through repository layer | ERROR |
Read CLAUDE.md for project-specific anti-patterns and blocked patterns. Flag any matches as ERROR.
CODE REVIEW
===========
File: [path]
Layer: [detected layer]
Language: [from CC_LANGUAGE]
STANDARDS
---------
[check]: [PASS/FAIL] [detail if failed]
ARCHITECTURE
------------
[check]: [PASS/FAIL] [detail if failed]
ANTI-PATTERNS
-------------
[check]: [PASS/FAIL] [detail if failed]
SUMMARY
=======
Category | Pass | Warn | Fail
-------------|------|------|-----
Standards | N | N | N
Architecture | N | N | N
Anti-patterns| N | N | N
VERDICT: [APPROVED | NEEDS_CHANGES]
/pre-commit -- Quick lint before staging/fitness -- Quality fitness scoringCLAUDE.md -- Project standards reference