antigravity-claudekit/skills/ck-fix/SKILL.md

name, description

name	description
ck-fix	Fix bugs, errors, test failures, CI/CD issues with intelligent routing. Always activate before fixing any bug, error, type error, lint issue, log error, UI issue, or code problem.

ck-fix

Unified skill for fixing issues of any complexity with intelligent routing and autonomous or human-in-the-loop workflow modes.

When to Use

• Any bug, error, or unexpected behavior requiring a fix
• Test suite failures (unit, integration, E2E)
• CI/CD pipeline failures
• TypeScript type errors or lint issues
• UI visual regressions
• Log errors from production or staging

Don't Use When

• The issue needs research and architectural planning first — use ck-planning
• You just want to understand the bug without fixing it — use ck-debug
• The task is a new feature, not a fix

Arguments

• --auto - Autonomous mode (default)
• --review - Human-in-the-loop review mode
• --quick - Fast mode for trivial bugs
• --parallel - Route to parallel agents per independent issue

Workflow

Step 1: Mode Selection

If no auto keyword in request, ask user to select workflow mode:

Option	Recommend When	Behavior
Autonomous (default)	Simple/moderate issues	Auto-approve if score >= 9.5 & 0 critical
Human-in-the-loop	Critical/production code	Pause for approval at each step
Quick	Type errors, lint, trivial bugs	Fast debug → fix → review cycle

Step 2: Debug

• Activate ck-debug
• Guess all possible root causes
• Dispatch multiple exploration agents in parallel to verify each hypothesis
• Create report with all findings

Step 3: Complexity Assessment & Fix

Level	Indicators	Workflow
Simple	Single file, clear error, type/lint	Quick: debug → fix → review
Moderate	Multi-file, root cause unclear	Standard: full pipeline
Complex	System-wide, architecture impact	Deep: research + brainstorm + plan
Parallel	2+ independent issues OR --parallel flag	Parallel developer agents

Step 4: Fix Verification & Prevent Future Issues

• Read and analyze all implemented changes
• Dispatch multiple exploration agents to find possibly related code
• Ensure fixes don't break other parts of the codebase
• Add comprehensive validation to prevent recurrence

Step 5: Finalize (MANDATORY — never skip)

1. Report summary: confidence score, changes, files modified
2. Update ./docs if changes warrant it
3. Ask user if they want to commit via ck-git

Skill Activation Matrix

Always activate: ck-debug (all workflows) Conditional: ck-sequential-thinking, ck-brainstorm, ck-research Parallel agents: Multiple exploration agents for scouting; verification agents for confirming fixes

Output Format

Step 0: [Mode] selected - [Complexity] detected
Step 1: Root cause identified - [summary]
Step 2: Fix implemented - [N] files changed
Step 3: Tests [X/X passed]
Step 4: Review [score]/10 - [status]
Step 5: Complete - [action taken]

Specialized Scenarios

• CI/CD failures: Fetch GitHub Actions logs via gh command, analyze pipeline steps
• Log errors: Parse application logs, correlate with code changes
• Test failures: Run test suite, analyze failure output, bisect if needed
• Type errors: Check TypeScript compiler output, trace type mismatches
• UI issues: Use visual analysis tools, compare with reference screenshots