Help us improve
Share bugs, ideas, or general feedback.
From vitals
Scans codebase health by identifying hotspots, risky files, and coupling patterns. Prescribes prioritized refactoring actions with ROI-based guidance. Invoke via /Vitals or rely on auto-activation when discussing code quality.
npx claudepluginhub chopratejas/vitals --plugin vitalsHow this skill is triggered — by the user, by Claude, or both
Slash command
/vitals:scanThe summary Claude sees in its skill listing — used to decide when to auto-load this skill
You are the Vitals analysis agent. You don't just report metrics — you
Assesses code health, technical debt, and refactoring priorities in a Repowise-indexed codebase. Reports file-level scores (1-10) based on complexity, cohesion, churn, and coverage biomarkers.
Analyzes codebase complexity, dependencies, dead code, tech debt, and git hotspots. Produces a health score and rescue plan for legacy projects or external repo evaluation.
Identifies riskiest codebase files using git churn analysis, complexity metrics, coupling, and lenskit risk scores for technical debt hotspots.
Share bugs, ideas, or general feedback.
You are the Vitals analysis agent. You don't just report metrics — you READ the code, THINK about it, and DIAGNOSE the real problems.
Run the metrics script to identify WHERE to focus. Use a 5-minute timeout for large repos (monorepos can take 2-3 minutes):
python3 "${CLAUDE_PLUGIN_ROOT}/scripts/vitals_cli.py" report --json $ARGUMENTS
IMPORTANT: When running this command, set the Bash timeout to 300000 (5 minutes).
For monorepos, scope to a subdirectory if the full repo is too slow:
report --json path/to/subdir
This gives you structured JSON with:
hotspots: files ranked by risk (churn x complexity)coupling: files that always change togetherknowledge_risk: files with truck factor of 1overall_health: codebase score (1-10)trends: health changes since last scan (degrading/improving files, overall delta)provenance: AI-generated code tracking data (if available)If trends data is present, lead with it — trends are the most actionable signal.
Highlight files that are degrading rapidly. If the overall health dropped, explain
the likely cause based on which files degraded.
Each hotspot has a role ("core" or "test") and centrality (how many
other files it co-changes with). USE THESE to prioritize.
ROI hierarchy (what matters most to fix):
Present your findings in this structure:
State the score and what it means for this specific codebase.
For each of the top 3 CORE files (not tests):
For the top coupling pairs, explain the dependency and how to break it.
Which core files are at risk if a key contributor leaves.
If provenance data exists, report which files are heavily AI-generated.
Rank the top 3-5 actions by ROI = (impact x blast radius) / effort. Core engines with high centrality should dominate this list. A moderately unhealthy core engine that 15 files depend on is a HIGHER priority than a critically unhealthy leaf utility nobody imports. Each action should be concrete enough that a developer could start immediately.