calebevans/GitHub Actions Failure Analysis
AI-powered analysis of GitHub Actions workflow failures
View on GitHubTrust Signals
- Scorecard Score
- not yet scored
- Maintenance Recency
- Stale
- License
- None
Inputs
| name | description | required | default |
|---|---|---|---|
| github-token | GitHub token for API access | yes | — |
| run-id | Workflow run ID to analyze (defaults to current run) | no | — |
| pr-number | Manual PR number override (for testing merged/closed PRs) | no | — |
| llm-provider | LLM provider (openai, anthropic, gemini, ollama) | yes | — |
| llm-model | LLM model name | yes | — |
| llm-api-key | LLM API key | yes | — |
| llm-base-url | Custom LLM API base URL | no | — |
| cordon-device | Device for cordon (cpu/cuda/mps) | — | cpu |
| cordon-backend | cordon embedding backend | — | sentence-transformers |
| cordon-model-name | cordon embedding model name | — | all-MiniLM-L6-v2 |
| cordon-api-key | API key for remote cordon embeddings | no | — |
| cordon-batch-size | Batch size for cordon embedding generation | — | 32 |
| post-pr-comment | Post analysis as PR comment | — | false |
| artifact-patterns | Comma-separated glob patterns for additional artifacts | no | — |
| ignored-jobs | Comma-separated job names to ignore | no | — |
| ignored-steps | Comma-separated step names to ignore | no | — |
| analyze-pr-context | Analyze failures in context of PR changes (true/false) | no | true |
| pr-context-token-budget | Percentage of context to allocate for PR diffs (0-50) | no | 20 |
Outputs
| name | description |
|---|---|
| summary | Analysis summary |
| category | Failure category |
| report-path | Path to full JSON report |