Operational Documentation Generator

You are generating operational documentation for a mission-critical TypeScript project. This skill consolidates multiple documentation types into a single workflow. Unlike /project-docs (which handles standard project files like README, LICENSE, CONTRIBUTING), this skill focuses on operational and technical documentation.

Available Document Types

Instructions

1. Determine document types

If $ARGUMENTS is provided, parse it as a comma-separated list of keys. Otherwise, ask the user which documents to generate using a multi-select question.

2. Load context

• •Read package.json for project name, dependencies, scripts
• •Read CLAUDE.md and README.md for project description
• •Scan src/ directory structure to understand module layout
• •Read .claude/docs/TypeScript Coding Standard for Mission-Critical Systems.md if present

Type: runbooks — Operational Runbooks

1. •

   Ask the user for:

   • •Service name
   • •Known dependencies (databases, caches, external APIs, message queues)
   • •Common operational scenarios they want documented (or use defaults)
2. •

   Create docs/runbooks/ directory with:

   docs/runbooks/TEMPLATE.md: Blank runbook template

   markdown

   # RB-NNN: <Title>
   
   **Severity**: SEV1 | SEV2 | SEV3 | SEV4
   **Last reviewed**: YYYY-MM-DD
   **Owner**: <team/person>
   
   ## Trigger Conditions
   <What alert or symptom initiates this runbook>
   
   ## Prerequisites
   - Required access: ...
   - Required tools: ...
   - Relevant dashboards: ...
   
   ## Diagnostic Steps
   1. ...
   2. ...
   
   ## Remediation Steps
   1. ...
   2. ...
   
   ## Rollback / Undo
   1. ...
   
   ## Escalation
   - When to escalate: ...
   - Contact: ...
   
   ## Post-Incident Verification
   1. Confirm the issue is resolved: ...
   2. Monitor for recurrence: ...
   
   ## Related Runbooks
   - [RB-NNN: Related Title](./related.md)
   

   Default runbooks (generate these unless the user specifies otherwise):

   • •high-latency.md — Diagnosing and resolving slow response times
   • •database-connection-failure.md — Handling database connectivity issues
   • •out-of-memory.md — Diagnosing memory leaks and OOM kills
   • •deployment-rollback.md — Rolling back a bad deployment
   • •dependency-outage.md — Handling upstream service outages
   • •certificate-expiry.md — Rotating expired TLS certificates

   docs/runbooks/README.md: Index table of all runbooks with ID, title, severity, and trigger.

Type: incident-playbook — Incident Response Playbook

1. •

   Ask the user for:

   • •Team/organization name
   • •On-call rotation tool (PagerDuty, OpsGenie, or none)
   • •Communication channels (Slack workspace, status page URL)
   • •Escalation contacts (or use placeholders)
2. •

   Create docs/incidents/ directory with:

   docs/incidents/playbook.md:

   • •Severity definitions (SEV1-SEV4) with impact criteria, response time targets, and communication cadence
   • •Roles: Incident Commander, Communications Lead, Subject Matter Expert, Scribe — each with responsibilities
   • •Incident lifecycle: Detection → Triage → Mitigation → Resolution → Post-mortem
   • •Communication templates: Slack incident channel message, status page update, stakeholder email (one per severity level)
   • •Escalation matrix with contact placeholders

   docs/incidents/post-mortem-template.md (blameless, modeled on Google SRE):

   • •Incident summary (what happened, duration, impact)
   • •Timeline (with timestamps)
   • •Root cause analysis (5 Whys method)
   • •Impact assessment (users affected, data loss, revenue impact)
   • •What went well / What went poorly / Where we got lucky
   • •Action items (owner, priority, due date)
   • •Lessons learned

   docs/incidents/severity-matrix.md: Detailed severity classification with examples.

   .github/ISSUE_TEMPLATE/incident.yml: GitHub Issue template for tracking incidents with structured fields.

Type: architecture — Architecture Diagrams (C4 Model)

1. •

   Analyze the project:

   • •Scan src/ directory structure — each subdirectory is a module
   • •Parse import statements across all .ts files to build a dependency graph
   • •Identify external dependencies from package.json (databases, HTTP clients, message queues)
   • •Read package.json description and README.md for system context
2. •

   Generate C4 diagrams in Mermaid syntax (renders natively in GitHub):

   docs/architecture/context.md — System Context diagram:

   • •The system as a box in the center
   • •External actors (users, external APIs, databases) around it
   • •Arrows showing interactions

   docs/architecture/components.md — Component diagram:

   • •Each src/ subdirectory as a component box
   • •Import relationships as arrows (with direction)
   • •External dependencies grouped separately
   • •Flag circular dependencies with a warning

   docs/architecture/dependencies.md — Module dependency graph:

   • •Text-based tree showing which modules import which
   • •Import counts per module
   • •Circular dependency detection

   docs/architecture/README.md: Index with descriptions and links.

3. •

   All diagrams use Mermaid code blocks:

   markdown

   ```mermaid
   graph TD
     A[Module A] --> B[Module B]
     B --> C[Module C]
   ```
   

Type: api-docs — API Documentation Setup

1. •

   Generate TypeDoc configuration (typedoc.json):

   json

   {
     "entryPoints": ["src/index.ts"],
     "out": "docs/api",
     "plugin": ["typedoc-plugin-markdown"],
     "exclude": ["**/*.test.ts", "**/*.spec.ts"],
     "readme": "none"
   }
   

2. •

   Add npm script: "docs:api": "typedoc" to package.json

3. •

   TSDoc completeness scan: Analyze all exported functions and types. For each public API, check for:

   • •Missing @param annotations
   • •Missing @returns annotation
   • •Missing @throws annotation (Rule 10.1)
   • •Missing usage examples for complex functions Report a TSDoc coverage percentage.
4. •

   If HTTP framework detected (Express, Fastify, Hono imports in deps):

   • •Ask if the user wants OpenAPI documentation
   • •If yes, generate an openapi.yaml skeleton with paths inferred from route definitions
   • •Suggest zod-to-openapi for automatic schema generation from existing Zod schemas
   • •List install commands: npm install --save-exact -D typedoc typedoc-plugin-markdown
5. •

   Optionally generate a CI job for auto-publishing docs to GitHub Pages on push to main.

Type: license-report — Third-Party License Attribution

1. •

   Scan dependencies: For each production dependency in package.json, determine its license by:

   • •Reading node_modules/<pkg>/package.json license field (if node_modules exists)
   • •Or running npm view <pkg> license
2. •

   Generate THIRD_PARTY_LICENSES.md:

   markdown

   # Third-Party Licenses
   
   This project uses the following third-party packages:
   
   | Package | Version | License | Repository |
   |---------|---------|---------|------------|
   | zod     | 3.22.4  | MIT     | https://github.com/colinhacks/zod |
   | ...     | ...     | ...     | ... |
   

3. •

   Generate or update license-policy.json:

   json

   {
     "allowed": ["MIT", "Apache-2.0", "ISC", "BSD-2-Clause", "BSD-3-Clause", "0BSD", "CC0-1.0"],
     "blocked": ["GPL-2.0-only", "GPL-3.0-only", "AGPL-3.0-only", "SSPL-1.0"],
     "requireReview": ["LGPL-2.1-only", "LGPL-3.0-only", "MPL-2.0", "CC-BY-4.0"]
   }
   

4. •

   Flag violations: Report any dependency with a blocked, unknown, or missing license.

5. •

   Suggest CI integration: Add a license-check step to the CI pipeline that validates all dependencies against the policy.

Summary

After generating all selected documents:

1. •List all created files
2. •List any install commands needed (TypeDoc, license-checker)
3. •Note any manual steps required (filling in placeholders, reviewing runbooks)