CI/CD Pipeline Polling

CI/​CD Pipeline Polling

This skill coordinates background monitoring of CI/​CD pipelines. When triggered, launch polling tasks that run independently and report back when builds complete.

When to Use

Trigger Keywords:

Automatic Triggers (PROACTIVE):

1. After git push to a branch with open PR/​MR
2. After creating PR with /​task:run github-pr
3. After creating MR with /​task:run gitlab-mr
4. When user mentions they're waiting for CI

Available Tasks

GitHub Actions

/​task:run github-poll-workflow [run-id]

• Polls GitHub Actions workflow runs
• Reports job-by-job status on completion
• Shows failed job logs automatically

GitLab CI/​CD

/​task:run gitlab-poll-pipeline [pipeline-id]

• Polls GitLab CI/​CD pipelines
• Reports stage-by-stage job status
• Shows failed job traces automatically

Usage Patterns

Pattern 1: After Push

When user pushes code:

User: "push this to the feature branch"
[After git push completes]
Claude: "Pushed to feature-branch. I'll monitor the CI workflow."
[Launch polling task in background]

Pattern 2: Explicit Request

User: "wait for the build to finish"
Claude: "I'll monitor the pipeline and report back when it completes."
[Launch polling task]

Pattern 3: Status Check

User: "is the build passing?"
[Check current status]
Claude: "Pipeline #12345 is running (3/​5 jobs complete). Want me to monitor until it finishes?"
User: "yes"
[Launch polling task]

Pattern 4: After PR Creation

User: "/​task:run github-pr"
[After PR is created]
Claude: "PR #42 created. Monitoring CI checks..."
[Launch polling task automatically]

Response Templates

Starting Polling

**Monitoring CI/​CD**

I'm watching [GitHub workflow / GitLab pipeline] #[ID] for branch `[branch]`.

- **Polling interval**: 30 seconds
- **Timeout**: 60 minutes
- **Current status**: [pending/​running]

I'll report back when it completes.

Success Report

**CI/​CD Complete: SUCCESS**

[GitHub workflow / GitLab pipeline] #[ID] passed.

**Branch**: `[branch]`
**Duration**: [X minutes]
**Jobs**: [N/​N passed]

The build is green. Ready to merge.

Failure Report

**CI/​CD Complete: FAILED**

[GitHub workflow / GitLab pipeline] #[ID] failed.

**Branch**: `[branch]`
**Failed Jobs**:
- [job-name]: [brief error]

**Last 50 lines of failed job:**

[error log excerpt]

Would you like me to:
1. Investigate the failure
2. Show full job logs
3. Retry the pipeline

Platform Detection

Detect platform automatically:

# Check remote URL
REMOTE=$(git remote get-url origin 2>/​dev/​null)

if echo "$REMOTE" | grep -q "github"; then
  # Use GitHub polling
  /​task:run github-poll-workflow
elif echo "$REMOTE" | grep -q "gitlab"; then
  # Use GitLab polling
  /​task:run gitlab-poll-pipeline
else
  echo "Unknown CI platform. Please specify:"
  echo "- /​task:run github-poll-workflow [id]"
  echo "- /​task:run gitlab-poll-pipeline [id]"
fi

Configuration

Environment Variables

Adjusting Polling

# Fast polling (for quick builds)
INTERVAL=10 /​task:run github-poll-workflow

# Long timeout (for slow builds)
TIMEOUT=120 /​task:run gitlab-poll-pipeline

Integration with Other Skills

• git: After push operations, offer to monitor CI
• gitflow: After finishing features/​releases, monitor CI
• tasks: After PR/​MR creation, monitor CI
• promptcontext: Set CI context for repeated monitoring

Best Practices

1. Don't poll too fast: 30s interval avoids rate limits
2. Set reasonable timeouts: Most builds finish in <30 min
3. Report failures clearly: Show actionable error info
4. Offer next steps: After failure, suggest investigation
5. Run in background: Use context: fork for non-blocking monitoring