From blueprint
This skill should be used when the user asks to "review my code", "check for bugs", "check style", "review changes", "run code review", "check for issues", "check for performance issues", or needs code reviewed for style compliance, bugs, logic errors, or efficiency. Also use when the user mentions reviewing staged changes, checking code quality, or finding problems in code.
npx claudepluginhub urso/claudev --plugin blueprintThis skill is limited to using the following tools:
Comprehensive code review checking style guide compliance, bugs/logic errors, and performance inefficiencies. Runs specialized reviews in parallel, then validates findings for false positives.
Provides Ktor server patterns for routing DSL, plugins (auth, CORS, serialization), Koin DI, WebSockets, services, and testApplication testing.
Conducts multi-source web research with firecrawl and exa MCPs: searches, scrapes pages, synthesizes cited reports. For deep dives, competitive analysis, tech evaluations, or due diligence.
Provides demand forecasting, safety stock optimization, replenishment planning, and promotional lift estimation for multi-location retailers managing 300-800 SKUs.
Comprehensive code review checking style guide compliance, bugs/logic errors, and performance inefficiencies. Runs specialized reviews in parallel, then validates findings for false positives.
${CLAUDE_PLUGIN_ROOT}/scripts/list-rules.sh$ARGUMENTS
Parse for:
--story <story-file> for context (optional)--style-only to run only style review--bugs-only to run only bug review--efficiency-only to run only efficiency review!git diff --name-only HEAD
!git diff --name-only --cached
If files specified: Use those files.
If no files specified: Use the changed files from the pre-computed context above.
Filter out non-code files:
*.md, *.json, *.yaml, *.yml, *.txt, *.lock.git/, node_modules/, vendor/, dist/, build/If no code files to review, report that and exit.
If --story specified, read the story document for implementation context.
Read the relevant reference files for review instructions, then spawn sub-agents in parallel. Each agent receives the review instructions from the reference file, the file list, and story context if provided.
Unless a --*-only flag limits the scope, spawn all applicable agents.
Style review (skip if --bugs-only or --efficiency-only):
Task: Run style review
Model: sonnet
Read the style review instructions from references/style.md.
Review these files: [file list]
Story context: [if provided]
Follow the instructions to review each file and report issues.
Bug review (skip if --style-only or --efficiency-only):
Task: Run bug review
Model: opus
Read the bug review instructions from references/bugs.md.
Review these files: [file list]
Story context: [if provided]
Follow the instructions to review each file and report issues.
Efficiency review (skip if --style-only or --bugs-only):
Task: Run efficiency review
Model: sonnet
Read the efficiency review instructions from references/efficiency.md.
Review these files: [file list]
Story context: [if provided]
Follow the instructions to review each file and report issues.
Wait for all agents to complete, then merge findings grouped by severity:
If issues were found, read references/validate.md for validation instructions. Spawn validation sub-agents to check each issue category with appropriate model tiers (haiku for style, sonnet for bug warnings, opus for critical bugs).
## Code Review Summary
Reviewed X files (validated for false positives).
### Critical Bugs (N)
[validated error-level bug issues]
### Bug Warnings (N)
[validated warning-level bug issues]
### Style Issues (N)
[validated style issues]
### Efficiency Issues (N)
[validated efficiency issues]
### Clean Files
[files with no confirmed issues]
---
Initial findings: X issues
After validation: Y confirmed issues (Z filtered as false positives)
If validated issues reveal recurring patterns not yet documented in style guides, propose additions:
## Style Guide Proposal
**Guide**: [which style guide, e.g., go.md, common.md]
**Rule**: [short rule name]
**Description**: [what the rule captures]
**Rationale**: [why this matters, based on issues found]
Add to the style guide? (Run `/style-update` to add)
Only propose if the issue represents a generalizable, actionable pattern not already covered.
Detailed review instructions for each category:
references/style.md — Style guide compliance checks, output format, review standardsreferences/bugs.md — Bug/logic/security review categories, severity levelsreferences/efficiency.md — Performance review categories, context-dependent filteringreferences/validate.md — False positive validation with tiered model sub-agents/review-fix — automated review + fix loop/review-loop — simplify, fix, and review-fix cycle/develop-loop — full development workflow/review-code # Review all staged/unstaged code files
/review-code src/auth/ # Review specific directory
/review-code --story 0001-*-setup # Review with story context
/review-code --style-only # Only check style compliance (faster)
/review-code --bugs-only # Only check for bugs (opus)
/review-code --efficiency-only # Only check for inefficiencies