- Home
- Skills
- Dhughes
- Claude Marketplace
- Gcb Monitor
gcb-monitor_skill
- Go
0
GitHub Stars
1
Bundled Files
2 months ago
Catalog Refreshed
4 months ago
First Indexed
Readme & install
Copy the install command, review bundled files from the catalogue, and read any extended description pulled from the listing source.
Installation
Preview and clipboard use veilstrat where the catalogue uses aiagentskills.
npx veilstrat add skill dhughes/claude-marketplace --skill gcb-monitor- SKILL.md7.7 KB
Overview
This skill monitors Google Cloud Build executions and provides continuous status updates until builds complete. It uses the gcloud CLI to find builds, stream logs, and perform focused failure analysis. It reports progress, notifies on completion, and gives actionable remediation steps when builds fail.
How this skill works
It identifies the correct GCP project from context or explicit input, locates the target build (by ID, branch, trigger, or commit SHA), then polls gcloud builds describe every 60 seconds. While running it streams relevant logs, highlights step-level progress, and on failure extracts the failed step and error messages. It also integrates with GitHub PR checks to map PRs to Cloud Build runs.
When to use it
- After merging a PR to monitor the resulting CI build
- When a user asks “check build status” or “what happened with the build”
- Before or during deployments to staging or production
- When you want to use gcloud CLI to inspect build logs or results
- If a build is stuck in QUEUED or WORKING and you need step-level insight
Best practices
- Confirm or infer the correct GCP project before querying builds
- Prefer build ID when available; otherwise filter by branch, trigger, or commit SHA
- Poll every 60 seconds and report summary updates; give detailed logs on failures
- When diagnosing failures, identify the failed step, extract error messages, and link to the console log URL
- Use human-friendly names (branch or trigger) for spoken notifications; avoid mentioning UUIDs
Example use cases
- Monitor the production build after a merge to main and get notified on success or failure
- Watch a staging build for a hotfix and stream failing test output to triage quickly
- Investigate a failed deployment step by retrieving step logs and recommended fixes
- Map a GitHub PR check to a Cloud Build run and follow its logs until completion
- Continuously poll a long-running build and report step completions and final status
FAQ
I will check gcloud auth list and prompt you to run gcloud auth login if no active account is found.
How long will monitoring continue?
Monitoring continues until the build reaches SUCCESS, FAILURE, CANCELLED, or TIMEOUT. I keep retrying transient API errors and never stop for temporary failures.
Can you show logs for a specific step?
Yes. I use gcloud builds log BUILD_ID --stream and target the time window for the failed step to extract the relevant output and error messages.