Marketplace - GitHub CICD Recovery Assistant Agent
This agent automates CI/CD failure triage by analyzing failed GitHub Actions workflows, classifying failures as transient or code-related, and taking appropriate follow-up actions. It can automatically re-run workflows and create GitHub issues for failures requiring developer attention.
Solution Details
- Difficulty Level
- BEGINNER
- Solution Type
- Designer Agent Accelerator
- Author
- Nexus Transformation
- Published on
- May 7, 2026
- Last updated on
- May 7, 2026
Get Started
Get StartedHow it works
GitHub CICD Recovery Assistant Agent automates the analysis of failed GitHub Actions workflows by inspecting logs and error patterns to determine whether failures are transient or code-related. Based on the classification, it can re-run workflows with approval or create GitHub issues for failures requiring developer attention
Features
- Pulls recent failed GitHub Actions workflow runs by repository, branch, status, or timeframe. Helps engineers quickly identify pipeline failures without manually reviewing GitHub Actions dashboards.
- Reads workflow job logs to identify the exact failed step, extract error messages, and highlight issues such as test failures, dependency errors, or authentication problems.
- Automatically determines whether a workflow failure is caused by transient infrastructure issues or genuine code defects that require developer intervention.
- Recommends the best recovery action based on failure type, including re-running failed jobs, escalating issues, or creating tickets for engineering teams
- Re-runs only failed GitHub Actions jobs instead of restarting the entire workflow, reducing unnecessary compute usage and speeding up recovery.
- Creates structured GitHub issues for persistent failures with relevant details such as run ID, branch, workflow name, logs, and remediation recommendations.
- Detects recurring workflow failures across multiple runs and identifies repeated patterns such as flaky dependencies or unstable infrastructure components.
- Identifies which developer, branch, pull request, or commit triggered a failed workflow to help teams resolve issues faster.
- Generates workflow health summaries showing failure trends, success rates, and the most commonly failing pipelines over time.
Verified Partner/Vendor Supported: This accelerator is fully supported and managed by the Partner/Vendor. For any support related inquiries or assistance including troubleshooting, guidance and/or updates to the solution, please contact the Partner/Vendor directly.
Applications Required
- GitHub