How does WarpFix detect CI failures?

WarpFix integrates as a GitHub App. When a workflow fails, GitHub sends a webhook and WarpFix starts analyzing the failure logs immediately.

Is my code safe with WarpFix?

WarpFix only reads logs and file contents needed for repair. All patches are validated in isolated sandboxes. The GitHub App uses minimal permissions.

What errors can WarpFix fix?

Build errors, test failures, lint issues, type errors, dependency problems, runtime crashes, and configuration bugs.

How does fingerprint learning work?

Each error is normalized and hashed into a unique fingerprint. When the same pattern recurs across any repository in your org, the proven fix is reused instantly with high confidence — reducing repair time by 37%.

What if a patch has low confidence?

Patches below 40 confidence are flagged for manual review and not opened as PRs. Scores between 40-70 get a 'review suggested' label. Only high-confidence patches are auto-merged.

Can I use WarpFix without Warp terminal?

Yes. The web dashboard and GitHub App work independently. Terminal commands are optional — you can use WarpFix entirely through the web interface.

How is WarpFix different from CodeRabbit?

CodeRabbit reviews code but does not fix it. WarpFix reviews AND auto-repairs CI failures. WarpFix also includes predictive CI failure prevention, security auto-patching, dead code detection, fingerprint learning, and sandbox validation — features CodeRabbit does not offer.

What CI/CD platforms does WarpFix support?

WarpFix currently supports GitHub Actions with webhook integration. Support for GitLab CI, CircleCI, and Jenkins is on the roadmap.

Settings

Configure how WarpFix reviews and repairs your code

Review Profile

Assertive

Thorough — flags all issues including style

Chill

Lenient — only critical issues, no nitpicks

Automation

Auto Review PRs

Automatically review every new PR

Auto Repair CI

Automatically fix CI failures

.warpfix.yaml

Add to your repo root to customize per-repo behavior

# .warpfix.yaml
review:
  profile: assertive
  auto_review: true
  path_instructions:
    - glob: "src/auth/**"
      instruction: "Check for security vulnerabilities"
    - glob: "**/*.test.js"
      instruction: "Verify test coverage"
  ignore_paths:
    - "dist/**"
    - "*.min.js"

repair:
  auto_fix: true
  min_confidence: 60
  skip_branches:
    - "release/*"

analysis:
  dead_code: true
  security_scan: true
  predictive_ci: true

quality_gates:
  - name: "No console.log in production"
    rule: "No console.log in src/ files"

chat:
  enabled: true