Procedure to add or change a GitHub Actions workflow: pin actions, set token permissions, cache, reuse workflows, and debug a failed run. Use when you edit files under .github/workflows/.
Procedure to add or change a GitLab CI pipeline: validate .gitlab-ci.yml, structure jobs with rules/needs/include, cache, and debug a failed job. Use when you edit .gitlab-ci.yml or files under .gitlab/ci/.
Procedure to verify a PostgreSQL or MySQL backup by restoring it into a scratch instance, and to run a point-in-time restore. Use when you check backups, set up backup jobs, or must restore data.
Procedure to generate an SBOM, scan it for vulnerabilities, and sign and verify container images and attestations with cosign. Use when you add SBOM, image signing, or provenance to a pipeline or verify an image.
Procedure to scan a repository and its history for secrets with gitleaks, triage hits, rotate a leaked secret, and add the scan to pre-commit and CI. Use when a secret leaks or when you set up secret scanning.
Procedure to detect and resolve drift between Terraform/OpenTofu code, state, and real infrastructure with plan -detailed-exitcode, import, and refresh-only. Use when a plan shows unexpected changes or when you set up scheduled drift checks.
Procedure to create or change a Terraform/OpenTofu module: inputs and outputs, version pins, fmt/validate/tflint/test, and safe renames with moved blocks. Use when you edit *.tf files of a module or move resources between modules.
Procedure to change a Helm chart: edit values and templates, lint, render, validate against the cluster schema, diff against the live release, and bump the chart version. Use when you edit files under charts/ or a Chart.yaml.
Decision procedure to find why a Kubernetes pod is Pending, CrashLoopBackOff, OOMKilled, ImagePullBackOff, not Ready, or unreachable, with read-only kubectl commands. Use when a workload does not start or does not serve traffic.
Procedure to add or change a Prometheus alerting or recording rule and its Alertmanager route, with promtool unit tests and amtool route checks. Use when you edit Prometheus rule files, PrometheusRule objects, or alertmanager.yml.
Procedure to define an SLO for a service and implement it in Prometheus: choose the SLI, write recording rules, and add multi-window multi-burn-rate alerts. Use when you add or change an SLO, an error budget, or burn-rate alerts.
Procedure to run a production incident: declare and set severity, gather evidence, mitigate before root cause, keep a timeline, communicate, and hand off. Use when a service is down or degraded or an alert pages.
Procedure to add or change an HTTP (OpenAPI) or gRPC (protobuf) API contract: edit the spec first, check breaking changes with oasdiff or buf, version and deprecate, and add contract tests. Use when you add or change an endpoint, request or response schema,…
Procedure to audit dependencies of a Python, Go, Java, or Node project: scan, check reachability, classify each finding, and fix or record an exception. Use when you add or update dependencies, triage a vulnerability alert, or prepare a release.
Analyze one or many local Git repositories with agent-guides tools/repo-scan - activity class, metrics, dependency graph between repositories (libraries, HTTP, Kafka, gRPC, CI, deploy) with evidence, agent-guides applicability, quality gate candidates, and an…
Shows or switches the Agentic role profile (lean, team, or a custom id) of this project; use when the user asks which roles are installed or asks to switch to lean, team, or another profile.