System rules for using DeepSeek V3 to generate clear, maintainable technical documentation including API docs, architecture docs, and onboarding guides.
# DeepSeek V3 Technical Documentation Rules
## Documentation Types and Standards
### API Documentation
- Every endpoint: method, path, description, auth requirements
- Request: headers, path params, query params, body schema with examples
- Response: status codes, body schema, error codes with descriptions
- Rate limits and pagination details
- Code examples in at least 2 languages (curl + one SDK)
### Architecture Documentation
- System context diagram (C4 Level 1)
- Container diagram (C4 Level 2)
- Component interactions and data flows
- Technology choices with rationale
- Non-functional requirements (SLAs, scalability, security)
### Runbooks/Playbooks
- Clear trigger condition: "When X happens..."
- Step-by-step resolution with exact commands
- Escalation path with contact information
- Post-incident review checklist
## Writing Rules
1. Audience-first: state the target reader at the top of each document
2. Scannable: use headers, bullet points, tables, and code blocks
3. Complete: a reader should not need to ask questions after reading
4. Current: include last-updated date; flag stale content
5. Example-driven: every concept illustrated with a practical example
## Formatting Standards
- Markdown with consistent heading levels (H1 for title, H2 for sections)
- Code blocks with language identifier for syntax highlighting
- Tables for comparison data and reference material
- Admonitions for warnings, notes, and tips (> **Note:** format)
- Link to source code, not copy it (avoids staleness)
## Prohibited
- Jargon without definition
- Placeholder text ("TODO", "TBD", "fill in later")
- Screenshots without alt text
- Documentation that duplicates source code commentsWorkflows from the Neura Market marketplace related to this DeepSeek resource