CI Fix Executor

Diagnose and fix CI failures so PRs can pass without manual debugging.

JTBD (Jobs To Be Done)

Primary Job

When CI has failed on a PR and I need the failure diagnosed and fixed, I want automated diagnosis, targeted fix generation, and local validation, so the PR can pass CI without me manually parsing log output.

Secondary Jobs

• When a CI failure is blocking a merge and I need a quick targeted fix rather than a full investigation.
• When CI logs are long and noisy and I need the root cause extracted and acted on.
• When qa-critic identified a flaky test pattern and I need to fix the underlying cause.

This Skill Is For

• CI failures on PRs (GitHub Actions, CircleCI, or generic CI systems)
• 5 failure categories: lint errors, test failures, build failures, deploy failures, dependency conflicts
• Generating minimal targeted fixes (not refactors)
• Local validation before re-triggering CI

This Skill Is NOT For

• Writing new tests from scratch; use test-planner
• Investigating flaky tests without a current CI failure; use qa-critic
• Performance optimization; use perf-critic
• Refactoring code that happens to be near the failure

Resolution Paths

User's Situation	What Happens	What They Leave With
Lint failure on PR	Executor parses lint output, identifies violations, applies auto-fix or generates targeted changes	A minimal diff fixing lint violations
Test failure on PR	Executor identifies the failing assertion, determines if test or implementation is wrong, generates targeted fix	A fix for the root cause (implementation fix or test correction with justification)
Build failure	Executor parses build output, identifies missing imports/config/deps, generates fix	A buildable state with the minimum change needed
Deploy failure	Executor identifies env/config/permission issues from deploy logs	Config fix or clear escalation with diagnosis
Dependency conflict	Executor resolves version constraints, updates lockfile	Resolved dependency tree

When to Escalate

• If the same fix approach fails twice, STOP and escalate to the user with full diagnosis
• If the failure appears to be a fundamental architecture issue, escalate to test-planner
• If the test is wrong but the behavior change was intentional, escalate to the user for confirmation

Paired With

• test-planner (optional upstream): provides testing strategy context
• qa-critic (downstream): reviews the fix for correctness and test health impact