Setup & Installation
Or with the ClawHub CLI, for registry-managed skill folders outside a full OpenClaw workspace:
What This Skill Does
Autonomous agent that scans TypeScript codebases for error handling anti-patterns such as empty catch blocks, swallowed promises, and unlogged errors. Operates in five modes: scan, review, auto, resume, and report. Session state persists across runs so large codebases can be fixed incrementally.
Combines static detection, templated fixes, and an approval workflow in one session-aware tool, replacing the manual grep-fix-grep cycle with tracked, resumable progress.
When to use it
- Finding silent catch blocks before a production incident
- Batch-fixing low-severity error handling issues before a code review
- Interactively reviewing error handling on critical database or auth paths
- Resuming a multi-session anti-pattern audit after switching branches
- Checking fix progress and remaining issue counts at any point in a refactor
Example Workflow
Here's how your AI assistant might use this skill in practice.
User asks: scan my src/ directory for error handling anti-patterns
- 1Runs `bunx antipattern-czar --src src` and captures output
- 2Checks project root for existing `.anti-pattern-state.json` and asks whether to resume
- 3Classifies each detected issue by severity and whether it sits on a critical path
- 4Saves full session state to `.anti-pattern-state.json`
- 5Displays a summary table grouped by severity with issue counts per file
Session saved with 34 issues across 12 files; 3 critical-path, 11 high, 20 medium — ready for REVIEW or AUTO mode