kud/claude-plugins/build-watch

Step 1 — Resolve the job and parameters

If the user provided a job name (e.g. /build-watch my-service), use it directly.

If the user provided parameters (e.g. /build-watch my-service BRANCH=main ENV=staging), parse them as key-value pairs.

If no job name was given, ask: "Which Jenkins job do you want to trigger?"

Step 2 — Trigger the build

Call trigger_build with the job name and any provided parameters.

Confirm the build was queued and note the queue ID if returned.

Step 3 — Poll until complete

Call get_job_status every 15 seconds to check whether the build has started and what its current state is.

Once the build number is known (it may take a few seconds to move from queue to running), switch to get_build_status with the specific build number.

Report status updates as they happen:

• QUEUED → "Build is queued..."
• RUNNING → "Build #N is running (started <time>)..."
• SUCCESS → "✅ Build #N passed"
• FAILURE → "❌ Build #N failed"
• ABORTED → "⚠️ Build #N was aborted"

Stop polling after 30 minutes. If the build hasn't finished by then, say so and provide the build URL.

Step 4 — Post-build summary

On completion, present:

### Build <job-name> #<N> — <STATUS>
 
**Duration**: <duration>
**Triggered by**: <cause>
**Branch**: <branch if available>

If the build failed, automatically run the ci-diagnose skill to explain the failure.

Step 5 — Offer next steps (optional)

Offer to:

• Fetch get_test_results if tests ran
• Fetch get_artifacts if the build produced artifacts
• Replay the build with replay_build if it failed due to a transient issue