Runbook

You are in AUTONOMOUS MODE. Do NOT ask questions. Analyze the project's infrastructure and operations artifacts to produce a comprehensive runbook.

INPUT: $ARGUMENTS

Accepted arguments:

• No arguments: generate a complete operations runbook.
• deployment: focus on deployment procedures only.
• incident: focus on incident response procedures only.
• database: focus on database maintenance procedures only.
• troubleshooting: focus on troubleshooting guides only.

============================================================ PHASE 1: INFRASTRUCTURE DISCOVERY

Step 1.1 -- Deployment Configuration

Scan for deployment artifacts:

Step 1.2 -- CI/CD Pipeline

Scan for: .github/workflows/*.yml, .gitlab-ci.yml, Jenkinsfile, .circleci/config.yml, bitbucket-pipelines.yml, cloudbuild.yaml, azure-pipelines.yml.

Read each pipeline file and extract build steps, deployment targets, environment variables, secrets referenced, and branch triggers.

Step 1.3 -- Database Configuration

Scan for: migration files (migrations/, prisma/migrations/, alembic/, db/migrate/), schema files (schema.prisma, models.py, *.entity.ts), seed files, database config (database.yml, knexfile.js), and backup scripts.

Step 1.4 -- Monitoring and Observability

Scan for: logging config (winston, pino, log4j), APM (newrelic, datadog, elastic-apm), metrics (prometheus, grafana), error tracking (sentry, bugsnag, rollbar), health check endpoints, and alerting rules.

Step 1.5 -- Environment Configuration

Scan for: .env.example, environment-specific configs, feature flags, and secrets management (vault, AWS SSM, sealed secrets).

============================================================ PHASE 2: RUNBOOK GENERATION

Generate docs/runbook.md with these sections. Only include sections relevant to detected infrastructure. Every procedure must have copy-pasteable commands.

The runbook MUST cover these areas (skip if not applicable to project):

1. Deployment -- Prerequisites, step-by-step deploy commands per environment, post-deploy verification checklist (health check, logs, smoke test, metrics).

2. Rollback -- Quick rollback procedure (< 5 min, redeploy previous version), full rollback with database revert if needed, verification checklist.

3. Scaling -- Scale up/down commands, auto-scaling config and thresholds.

4. Database Maintenance -- Run/rollback migrations, create/restore backups, common diagnostic queries (table sizes, active connections, slow queries).

5. Incident Response -- Severity levels (SEV1-4 with response times and examples), incident checklist (acknowledge, communicate, investigate, mitigate, resolve, postmortem), first-response diagnostic commands for logs, status, and resources.

6. Monitoring -- Dashboard links, key metrics with normal ranges, alert response guide (high CPU, memory, error rate, disk, health check failures with actions).

7. Troubleshooting -- Common failure scenarios (app won't start, connection timeouts, performance degradation) with step-by-step diagnosis. Error-to-solution lookup table.

Format: use markdown headers, numbered steps with fenced bash blocks for commands, checklists for verification, and tables for reference data.

If a runbook already exists, read it first and preserve manually-written sections.

============================================================ PHASE 3: VALIDATE AND WRITE

For each command in the runbook:

• Verify the referenced tool/CLI exists in the project's dependencies or Dockerfile
• Verify paths and file references are correct
• Verify environment variable names match what the project uses

Write the runbook to docs/runbook.md.

============================================================ SELF-HEALING VALIDATION (max 2 iterations)

After producing documentation, validate completeness:

1. Verify all required sections are present and non-empty.
2. Verify internal cross-references and links resolve correctly.
3. Verify no placeholder text remains ("{TODO}", "[TBD]", "...", "etc.").
4. Verify code examples are syntactically valid.

IF VALIDATION FAILS:

• Identify which sections are incomplete or contain placeholders
• Re-generate only the deficient sections
• Repeat up to 2 iterations

============================================================ OUTPUT

Runbook Generated

Infrastructure Detected

• Deployment: [Docker / K8s / Serverless / etc.]
• CI/CD: [GitHub Actions / GitLab CI / etc.]
• Database: [PostgreSQL / MongoDB / etc.]
• Monitoring: [Datadog / Prometheus / etc.]
• Environments: [dev, staging, production]

Sections Generated

• Deployment procedures
• Rollback procedures
• Scaling instructions
• Database maintenance
• Incident response
• Monitoring and alerts
• Troubleshooting guide

File Written

• docs/runbook.md

============================================================ SELF-EVOLUTION TELEMETRY

After producing output, record execution metadata for the /evolve pipeline.

Check if a project memory directory exists:

• Look for the project path in ~/.claude/projects/
• If found, append to skill-telemetry.md in that memory directory

Entry format:

### /runbook — {{YYYY-MM-DD}}
- Outcome: {{SUCCESS | PARTIAL | FAILED}}
- Self-healed: {{yes — what was healed | no}}
- Iterations used: {{N}} / {{N max}}
- Bottleneck: {{phase that struggled or "none"}}
- Suggestion: {{one-line improvement idea for /evolve, or "none"}}

Only log if the memory directory exists. Skip silently if not found. Keep entries concise — /evolve will parse these for skill improvement signals.

============================================================ DO NOT

• Do NOT fabricate infrastructure details. Only document what is evident from config files.
• Do NOT include actual secrets, passwords, or API keys in the runbook.
• Do NOT include placeholder commands for infrastructure that does not exist.
• Do NOT generate a runbook for projects with no deployment configuration (e.g., a library or CLI tool). Report that no runbook is needed and suggest /readme.
• Do NOT include generic advice not specific to this project's stack.

NEXT STEPS:

After generating the runbook:

• "Run /document to check overall documentation health."
• "Run /diagram to generate infrastructure diagrams."
• "Run /devops to review and improve the deployment pipeline."