jjjermiah/code-reviewer
src/dot-agents/skills/code-reviewer/SKILL.md
Provides structured code review against plans and standards. Use when a feature is complete and needs validation, when reviewing code before merge, or when assessing quality and test coverage—e.g., "finished step X", "ready for review", "validate architecture", "check quality and tests".
$ install --global
skillsauthnpx skillsauth add jjjermiah/dotagents code-reviewerInstall this skill globally with one command. Works with Claude Code, Cursor, and Windsurf.
Security Scan Results
3 of 9 scanners reported clean
Some scanners were skipped, did not run, or reported a non-clean status. Review each row below.
3
Scanners Passed
9
Scanners in report
Clean
TrivyContainer and dependency vulnerability scanner
95%
Clean
SemgrepStatic code analysis for vulnerabilities
95%
Clean
mcp-scan (Snyk)Model Context Protocol security validation
95%
Skipped
Snyk (dep)Open source security scanning
50%
Skipped
Socket.devSupply chain security analysis
50%
Skipped
VirusTotalMulti-engine malware detection
50%
Skipped
CrowdStrikeAdvanced threat intelligence
50%
Skipped
OSV-ScannerOpen Source Vulnerability database check
50%
Skipped
OWASP Dep-Check
50%
Last scanned: Apr 25, 2026, 3:50 PM119.8s3 files scanned
SKILL.md
- name:
- code-reviewer
- description:
- Provides structured code review against plans and standards. Use when a feature is complete and needs validation, when reviewing code before merge, or when assessing quality and test coverage—e.g., "finished step X", "ready for review", "validate architecture", "check quality and tests".
Code Reviewer Skill
Purpose
You MUST complete a thorough code review before marking any work as done. Reviews without systematic checking always miss critical issues. Every time.
Validate completed work against original plans, identify deviations, assess code quality/maintainability/test coverage/security, and provide actionable recommendations with clear severity labels.
Review Procedure
Announce your review start: "Beginning code review using code-reviewer skill."
Complete ALL six steps. Skipping any step means missing issues:
- Plan alignment analysis (ALWAYS do this first)
- Compare the implementation against the original plan
- Identify deviations and assess impact
- Verify all planned functionality is present
- Reviews without plan comparison miss architectural drift. Every time.
- Code quality assessment (check EVERY file)
- Check correctness, error handling, and type safety
- Evaluate maintainability, naming, and project conventions
- Assess tests and coverage quality
- Look for security and performance risks
- Use references/checklists.md for domain-specific checklists
- Architecture and design review (NEVER skip for new features)
- Validate SOLID principles and separation of concerns
- Check integration with existing systems
- Assess scalability and extensibility
- Documentation and standards (ALWAYS check)
- Verify comments and documentation are accurate and necessary
- Confirm adherence to project standards
- Issue identification and recommendations
- Categorize findings and propose fixes
- Provide code examples when useful
- See references/examples.md for proper format
- Communication protocol
- We're reviewing together. I need your honest technical judgment.
- Ask for confirmation on significant plan deviations
- Recommend plan updates if the plan itself is flawed
- Acknowledge strengths before issues
After completing all steps: "Code review complete. [N] critical, [N] important, [N] suggestions."
Output Contract
You MUST return reviews in this exact structure. No exceptions.
-
Overview (1-3 sentences summarizing overall quality)
-
Findings (bulleted), each with:
- Severity: critical | important | suggestion
- Location: file path + line (if available)
- Rationale: why it matters
- Fix: specific recommendation
- Ordering: sort by severity (critical first), then by location
Reviews without severity labels = ambiguous priority. Reviews without fix recommendations = incomplete.
-
Tests / verification suggestions (always include if you found issues)
-
Final statement: "Review complete using code-reviewer skill."
Before finalizing: Verify you checked every file and followed all six review steps.
References (Load on Demand)
- references/checklists.md - Load when reviewing specific domains (API, database, frontend, security)
- references/examples.md - Load to see sample reviews demonstrating proper format and detail level
Related Skills
jjjermiah/skill-creator
development
VerifiedTrustedCommunity
Guides creation, validation, and packaging of AI agent skills with token-efficient design, progressive disclosure patterns, and YAML frontmatter best practices. Use when building new skills, updating existing skills, validating skill structure against standards, or packaging for distribution—e.g., "create skill", "validate SKILL.md", "package skill for sharing", "check description format".
1SKILL.mdUpdated Apr 25, 2026
jjjermiah/sdk-module-investigation
tools
VerifiedTrustedCommunity
Investigate and integrate weakly documented SDK/library modules (especially Azure SDKs) into code. Use when asked to "investigate module", "SDK", "client class", or when docs are missing/weak and you need to discover APIs, models, or usage patterns to implement integration.
1SKILL.mdUpdated Apr 25, 2026
jjjermiah/script-writer
tools
VerifiedTrustedCommunity
Write production-ready one-off scripts and automation utilities with proper error handling and safety patterns. Use when developing bash automation, Python CLI tools, shell scripts, system administration scripts, or command-line batch processing—e.g., "write a script to process files", "python one-liner for data conversion", "bash automation for backups", "shell script with error handling".
1SKILL.mdUpdated Apr 25, 2026
jjjermiah/r-testing
development
VerifiedTrustedCommunity
R package testing with testthat 3rd edition. Use when writing R tests, fixing failing tests, debugging errors, or reviewing coverage—e.g., "write testthat tests", "fix failing R tests", "snapshot testing", "test coverage".
1SKILL.mdUpdated Apr 25, 2026