| name | architecture |
| description | Use after PRD/domain model and before implementation. Designs system architecture, component boundaries, dataflow, integration points, security boundaries, logging, audit, failure handling, and demo-vs-production separation. |
| argument-hint | [system-or-feature] |
Architecture and Dataflow
Use when
- Need buildable architecture
- Need dataflow/security boundary
- External API/model usage must be documented
- Demo shortcuts must be separated from production
Do not use when
- No domain model exists; run domain-model first unless user asks for rough sketch
Inputs to inspect
- Domain model
- PRD
- Feature spec
- Data rules
- Existing repo stack
Files to create or update
docs/07_architecture.md- Updates to
rules/data-rules.md, rules/agent-rules.md, rules/compliance-rules.md
Operating sequence
- Identify runtime components.
- Assign responsibilities and ownership.
- Describe dataflow and trust boundaries.
- List external services with data sent/received, license, constraints, fallback.
- Define storage, retention, logging, audit.
- Define failure modes and fallbacks.
- Separate demo-mode mocks from production architecture.
Required Architecture Views
- Component view.
- Dataflow view.
- Agent execution view.
- Security/trust boundary view.
- Audit/logging view.
- Failure/fallback view.
External Boundary Rule
No external model/API/service may appear without:
- purpose,
- data sent,
- data received,
- permission basis or constraint,
- fallback,
- audit/log event.
Quality gates
- No “black box AI” component.
- Data leaving the system is explicitly described.
- Demo mocks are labeled.
Stop conditions
Stop and route backward when prerequisite artifacts are missing, when the requested action would erase user work, or when a high-risk workflow lacks evidence, approval, audit, or fail-closed handling.
Example invocation
/harness-engineering:architecture
Apply this skill to the current repository. Inspect existing artifacts first, update only what is necessary, and report files changed, decisions, assumptions, validation, risks, and next skill.
Shared completion and integrity
End with the completion report in ../../references/common-completion-report.md and follow ../../references/common-integrity-rules.md.
