Systematic and adaptable code review methodology using Language Server Protocol tools. Adapts to available LSP capabilities across programming languages. Use when user asks for code reviews, quality assessments, or specific analysis of codebases in any programming language.
Systematic and adaptable code review methodology using Language Server Protocol tools. Adapts to available LSP capabilities across programming languages. Use when user asks for code reviews, quality assessments, or specific analysis of codebases in any programming language.
Systematic 9-phase code review methodology using Language Server Protocol tools. Adapts to available LSP capabilities across programming languages.
Skill Methodology
Systematic 9-phase code review methodology using Language Server Protocol tools. Extends DEVELOPER profile with sequential analysis phases assessing code quality, architecture, and maintainability.
Phase 9 (Report) - Synthesize findings into actionable report with prioritized recommendations
[!IMPORTANT]
Review phases provide thorough investigation, tool-verified findings, and incremental validation for code quality assessment.
Systematic Approach
The phases must be completed sequentially:
DO complete each phase fully - Each phase delivers specific analysis findings
DO wait for approval - Confirm deliverables before proceeding to next phase
DO adapt approach - Adjust based on available LSP capabilities
DO NOT skip phases - Even if you think you have enough information
[!IMPORTANT]
Review quality improves through systematic tool usage and incremental validation.
Understanding Language Server Capabilities
The initial Project Discovery phase establishes which tools are available by calling get_server_capabilities tool. This response shows:
Which LSP capabilities are supported (supported: true/false)
Exact tool names mapped to each capability for this review
Tool-specific usage metadata and guidelines
[!IMPORTANT]
The server capabilities response is the authoritative source for determining which tools to use in subsequent phases.
Adapting the Review Process
The 9-phase methodology remains consistent across languages. Use the server capabilities response to determine which tools to invoke in each subsequent phase:
[!IMPORTANT]
Review systematically all phase instructions to understand the required execution steps.
Session Guidelines
Practical guidance for applying code review methodology consistently throughout review sessions.
Phase Execution Protocol
The primary protocol for large codebases is strictly one phase per turn to manage cognitive load, token consumption and maintain collaborative pacing. This protocol must be followed unless an explicit exception is granted.
Protocol (Always Followed after Each Phase)
Acknowledge completion with deliverable summary
Wait for user approval before proceeding
Report only what tools verify - no speculation
Efficiency Exception (Small Scopes)
For analysis scoped to a subset of the codebase (e.g., a small PR), analyze the complexity and estimated tool-call cost for the relevant phases. If this analysis determines the token cost is low and pacing can be accelerated, propose combining up to three related phases into a single turn for efficiency. The user must grant explicit approval before executing the combined phases.
During Each Phase
Best practices for executing each review phase effectively.
DO
Use all supported LSP tools systematically
Record exact numbers (don't estimate)
Note tool failures (explain what didn't work and why)
Take time for thorough analysis (patience over speed)
Verify findings with multiple tools when possible
Check tool metadata from capabilities response for usage guidance
DON'T
Skip phases unless an explicit exception is granted
Make statements without tool verification
Assume patterns without checking references
Rush to conclusions before completing analysis
Suggest fixes before finishing all phases
Quality Checklist
Before finalizing the review, verify:
✅ All 9 phases completed systematically
✅ Every claim backed by tool-verified data
✅ Exact file:line locations for all issues
✅ Critical issues have code examples for fixes
✅ Metrics table populated with real numbers
✅ Capability limitations documented
✅ Refactoring safety assessed
✅ Immediate actions clearly defined
Session Completion
The code review findings serve as actionable reference for improvements and refactoring work.
[!IMPORTANT]
If user requests a conversation log to document this code review session, the conversation-log skill provides developer-specific guidance for technical session documentation.