Triages static analysis findings by evaluating code context to distinguish genuine issues from irrelevant noise.
The Code Finding Assessor skill streamlines the development workflow by systematically triaging static analysis reports within the Claude Code environment. It empowers Claude to evaluate code complexity, duplication, coupling, and potential secrets by reading the actual source code to make informed judgment calls. By distinguishing between inherent complexity—such as CLI dispatchers or state machines—and actionable technical debt, the skill helps developers suppress 'noise' and focus on high-priority improvements. It provides a structured workflow for auditing, dismissing false positives, and generating comprehensive summaries of a codebase's health.
主な機能
01Deep-dive code inspection using file outlines and specific line range reading
02Context-aware triage of complexity, clones, coupling, and secrets findings
03Detailed summary reporting of 'Before vs. After' finding counts
04Bulk and individual dismissal of irrelevant findings with recorded rationale
056 GitHub stars
06Prioritized review workflow starting with critical security and complexity issues
ユースケース
01Filtering architectural boilerplate from duplication reports to focus on extractable logic
02Auditing security flags to separate genuine credentials from test fixtures and placeholders
03Automating the cleanup of static analysis 'noise' in large or legacy repositories
