cloud-resource-healthlisted
Install: claude install-skill sandbaseai/workbuddy-skill
# Cloud Resource Health
## Purpose and safety boundary
Diagnose a named cloud resource using evidence from the resource, its dependencies, recent changes, and the affected user path. This Skill is read-only by default. It does not reboot, resize, delete, change IAM/network policy, alter concurrency, or enable logging. Any remediation must have an owner, authorization, impact window, rollback/recovery plan, and post-change verification.
Do not assume that a provider status of “available” means the workload is healthy. Resolve the exact account/project, region, resource identifier, tenant scope, and time window before drawing conclusions. Redact account identifiers, customer data, tokens, private addresses, and raw log payloads.
## Inputs
Capture:
- exact error, affected user journey, first/last occurrence, frequency, and severity;
- provider, account/project, region, resource type and identifier, environment, and owner;
- recent deploys, config/IAM/network changes, scaling events, provider events, and incident timeline;
- relevant health metrics, logs, traces, dependency status, quotas, limits, and alert state;
- baseline comparison window and any concurrent changes or known data gaps.
If multiple resources match, stop and request a narrower scope. If telemetry is missing, report the blind spot and lower confidence rather than treating “no data” as “healthy.”
## Workflow
1. **Preserve and scope the symptom.** Create a short sanitized timeline and identify the failing