promptledger-verifiable-pro.../AGENTS.md

27 lines
1.3 KiB
Markdown

# AGENTS
Architecture overview for PromptLedger MVP
- Language/Stack: Python 3.8+ with standard library; packaging metadata via pyproject.toml
- Core modules:
- promptledger_verifiable_provenance_and_l: LocalProvenanceBlock, MerkleAuditLog, DeltaSync, Adapters
- Data model:
- LocalProvenanceBlock records a creative step with author, tool, action, metadata, license, timestamp, and a crypto-like signature
- Provenance storage:
- MerkleAuditLog keeps blocks and a Merkle root for tamper-evidence
- Synchronization: DeltaSync provides delta-exports/imports of provenance blocks for offline-first operation
- Adapters:
- BlenderAdapter, FigmaAdapter emit provenance blocks for their respective tools
- Governance: Draft governance scaffolding in code comments and API design; real RBAC and policy templates to extend later
- Testing:
- pytest-based tests under tests/
- Build/publish: python3 -m build; test.sh will run tests and build, per repository requirements
How to run locally:
- Install dependencies (only standard library for MVP): no extra deps required
- Run unit tests: pytest
- Build package: python3 -m build
Note: This is a minimal MVP to bootstrap the architecture. Future work will introduce real cryptographic PKI, policy engines, delta-sync optimizations, and cross-tool adapters.