Automated Code Fixer

An intelligent "IT guy" skill that automatically detects and fixes code issues with strict quality controls.

When This Skill Activates

• Test failures detected
• Linting errors reported
• Type-checking errors found
• Security vulnerabilities identified
• Build failures encountered
• CI/CD pipeline failures

Strict Quality Gates

CRITICAL: All fixes must pass these gates before being accepted:

Gate 1: Test Validation

• All existing tests must pass after fix
• New code must have corresponding tests
• Coverage cannot decrease below 70%

Gate 2: Linting Compliance

• Black formatting must pass
• Ruff linting with zero errors
• No new security warnings

Gate 3: Type Safety

• mypy type-checking must pass
• No untyped function signatures
• No Any type escapes

Gate 4: Security Check

• No hardcoded secrets
• No SQL injection vulnerabilities
• No path traversal issues
• Input validation on all user data

Gate 5: Architectural Compliance

• Follow layered architecture (Route -> Controller -> Service -> Model)
• Database changes require migrations
• Async/await for all DB operations

Fix Process

Step 1: Diagnose

# Run diagnostics
cd /home/user/Autonomous-Assignment-Program-Manager/backend
pytest --tb=short 2>&1 | head -50
ruff check app/ tests/
mypy app/ --python-version 3.11

Step 2: Analyze

• Identify root cause (not just symptoms)
• Check related files for context
• Review test expectations

Step 3: Fix

• Make minimal, focused changes
• Follow existing code patterns
• Add comments only where logic isn't self-evident

Step 4: Validate

# Must all pass before accepting fix
pytest --tb=short
ruff check app/ tests/
black --check app/ tests/
mypy app/ --python-version 3.11

Step 5: Report

Provide concise summary:

• What was broken
• Root cause
• What was fixed
• Tests added/modified

Escalation Rules

Escalate to human when:

1. Fix requires changing models/migrations
2. Fix affects ACGME compliance logic
3. Fix touches authentication/security code
4. Multiple interdependent failures
5. Unclear requirements or business logic
6. Fix would take >30 minutes

Rollback Protocol

If fix causes additional failures:

1. Immediately revert changes
2. Document what went wrong
3. Escalate to human with full context

Integration with Existing Commands

This skill works with the project's slash commands:

• /run-tests - Execute test suite
• /lint-fix - Auto-format and fix linting
• /health-check - System health validation
• /check-compliance - ACGME compliance verification

Example Fixes

Test Failure

pytest output: FAILED tests/test_swap_executor.py::test_execute_swap - AssertionError

Diagnosis: SwapExecutor.execute_swap() not awaiting async call
Fix: Added 'await' to database query on line 47
Validation: All tests pass, coverage maintained at 78%

Type Error

mypy output: app/services/schedule.py:23: error: Missing return type annotation

Diagnosis: Function missing type hints
Fix: Added -> Optional[Schedule] return type
Validation: mypy passes, no runtime changes

Security Issue

ruff output: S105 Possible hardcoded password in variable assignment

Diagnosis: Test file using hardcoded password
Fix: Replaced with environment variable via conftest fixture
Validation: No security warnings, tests pass

References

• See reference.md for detailed fix patterns
• See examples.md for common fix scenarios