# Recover CI Failures With Evidence-Based Fixes

Broken CI pipelines slow releases and hide root causes. This skill guides evidence-based failure analysis, repair planning, sandbox validation, and GitHub feedback.

## Install

```bash
npx skillstore add dnyoussef/cicd-intelligent-recovery
```

## Metadata

- Status: approved
- Slug: dnyoussef-cicd-intelligent-recovery
- Skillstore revision: r1
- Version status: missing
- Tree hash: 2f4b85bd85344d9d6224229290bfe2861663f8e771a7a1303d1209d9105e192b
- Author: DNYoussef
- GitHub username: DNYoussef
- License: MIT
- Repository: https://github.com/DNYoussef/ai-chrome-extension/tree/main/.claude/skills/cicd-intelligent-recovery
- Ref: 72d5025b022c77f7a51bdf5c1637c689c80e89d1
- Supported tools: Claude, Codex, Claude Code
- Audit status: complete
- Agent install advisory: confirmation\_required
- Manual install advisory: allowed
- Artifact signature: available
- Audit attestation: unavailable
- Human verification: not\_verified
- Risk factors: external\_commands, network, filesystem, env\_access
- Quality score: 38
- Quality tier: warning
- Public page: https://skillstore.pages.dev/skills/dnyoussef-cicd-intelligent-recovery
- Manifest: https://skillstore.pages.dev/api/skills/dnyoussef-cicd-intelligent-recovery/manifest

## Capabilities

- Collects GitHub Actions failure logs through gh CLI examples.
- Parses failure logs into structured .claude artifact files.
- Coordinates multi-agent root cause, dependency, and test validity analysis.
- Plans fixes with connascence context and validation checkpoints.
- Runs sandbox, theater, and differential validation workflows.
- Generates PR, status, and Loop 1 feedback report templates.

## Use Cases

- Triage Failed GitHub Actions Runs: Download failed workflow logs, parse failures, and prepare structured evidence for repair planning.
- Validate Test Repairs: Check whether fixes address real defects, avoid shallow test theater, and pass sandbox validation.
- Close the Feedback Loop: Summarize root causes, improvement metrics, and recurring failure patterns for future planning.

## Prompt Templates

### Review Recent CI Failures

```
Use this skill to review the latest failed GitHub Actions run. Summarize failing tests, likely causes, and required evidence.
```

### Build a Root Cause Report

```
Analyze the parsed CI failures. Separate primary root causes from cascaded failures. Include confidence levels and affected files.
```

### Plan Safe Repairs

```
Create a minimal repair plan for each root cause. Include affected files, validation steps, side effects, and rollback guidance.
```

### Run Full Recovery Workflow

```
Execute the complete recovery workflow with explicit approval before network calls, file writes, commits, pushes, or PR creation.
```

## Limitations

- Requires authenticated GitHub CLI and local permission to run commands.
- Assumes claude-flow alpha tooling is installed or available through npx.
- Does not guarantee safe autonomous code changes without human review.
- May expose repository data to external tools if prompts are run unchanged.

## Best Practices

- Review every shell, GitHub, and sandbox command before execution.
- Pin external tools and verify package integrity before using npx workflows.
- Redact secrets from logs and .claude artifacts before sharing reports.

## Anti Patterns

- Do not run the full workflow with repository write access unattended.
- Do not use git add . when artifacts may contain secrets or logs.
- Do not treat a 100 percent test pass as proof of production safety.

## Security Audit

- Audited at: 2026-07-06T12:07:01.161\+00:00
- Summary: The audit found many false positives from Markdown fences and JavaScript template literals, but the skill still contains real operational risk. Confirmed issues include shell command substitution, hidden artifact storage, filesystem writes, unpinned alpha CLI execution, GitHub publishing automation, and external sandbox uploads.

## Stats

- Views: 192
- Downloads: 7
- Favorites: 0
- Popularity score: 0