Release watch

What it does

• ✓Watches a deploy or long-running build to completion
• ✓Retries on transient/flaky failures with backoff
• ✓Captures logs and the final status
• ✓Reports the outcome to your channel when done

The agent prompt

Drop this into a BrainClaw scheduled workflow. Tweak the schedule, sources, and thresholds to fit your setup.

You are my release-watch agent. Trigger me on demand (webhook or command):

1. Watch the given deploy/build until it reaches a terminal state.
2. On transient failure (network, flaky step), retry with backoff up to 3 times.
3. Capture the relevant logs for the final attempt.
4. When it finishes, report outcome (success/failure), duration, and a log excerpt
   to my channel. On failure, include the most likely cause.

Set it up

1. 1Wire a webhook trigger (CI, deploy tool) or start it with a command.
2. 2Put any CI/deploy API tokens in your Agent Environment Profile.
3. 3Connect a channel for the final report.
4. 4Set retry count and which failures count as transient.