docs/ is the source of truth for the Agent Data Synthesis framework.
- DESIGN.md: system principles, bounded contexts, and contracts.
- BACKEND.md: backend service boundaries and execution model.
- DATA.md: canonical data schemas, lineage, and dataset outputs.
- SECURITY.md: sandboxing, trust boundaries, secrets, and external integrations.
- PRODUCT_SENSE.md: users, value model, success metrics, and non-goals.
- ROADMAP.md: staged development direction.
- PLANS.md: execution plan index and lifecycle state.
- design-docs/agent-data-synthesis-framework.md: full framework architecture.
- design-docs/architecture-explainers.md: detailed explanations of architecture and algorithm concepts such as the AWM environment model.
- product-specs/framework-mvp.md: MVP scope and acceptance criteria.
- references/agent-data-synthesis-pdf-analysis.md: structured analysis of
Agent-ๆฐๆฎๅๆ.pdf.
- generated/README.md: index for future generated schemas, diagrams, reports, and benchmark outputs.
- exec-plans/active/README.md: active plan bucket.
- exec-plans/active/0012-controlled-network-backed-environment-synthesis.md: current controlled network-backed environment synthesis implementation plan.
- exec-plans/active/0013-mcp-compatible-environment-tool-adapters.md: next MCP-compatible local environment/tool adapter plan.
- exec-plans/completed/README.md: completed plan bucket.
- exec-plans/completed/0011-provenance-licensing-and-sandbox-gates.md: latest completed source provenance, license eligibility, default-deny network policy, sandbox policy, source-event audit artifacts, and reporting slices.
- exec-plans/tech-debt/README.md: technical debt bucket.