Diagnose CI/CD pipeline failures

cicd-stage1-examplesubagentsetup L3★3

What it does

Orchestrate parallel CI/CD failure detection and analysis agents

Best for

Simultaneously discovering both failure symptoms (via failure-detector) and contextual causes (via context-analyzer) in CI/CD pipelines to handoff to Stage 2 root-cause agents.

Inputs

· git history 30 days
· GitHub Actions/GitLab CI/Jenkins pipeline logs

Outputs

· /tmp/cicd-pipeline-*/stage-1/detected-failures.json
· /tmp/cicd-pipeline-*/stage-1/analyzed-context.json

Requires

· GitHub Actions
· GitLab CI
· Jenkins
· Git

Preconditions

· Two subagents available: failure-detector, context-analyzer
· CI logs accessible
· 30-day git history available

Failure modes

· Subagent timeout if log volume exceeds agent capacity
· Missing correlation if events overlap
· Incomplete pattern detection if failure window too narrow

Trust signals

· Explicit parallel agent execution pattern
· Structured JSON output (detected-failures.json + analyzed-context.json)
· Stage 2 handoff design for root-cause analysis
· Timeline expectation: 2-5 minutes parallel execution