v0.1.0 - Secure & Verifiable Tool Invocation Flow

v0.1.0 - Secure & Verifiable Tool Invocation Flow

Initial marketplace-ready release candidate.

What this release includes

• CrewAI Flow wrapper for secure and verifiable tool invocation
• Reusable guarded_tool_call() module
• Deterministic canonical JSON and SHA-256 hashing
• Ed25519 signed execution receipts
• Exact-match policy checking
• Independent receipt validator with CLI support
• Optional demo-level replay detection
• Deterministic dataset metadata lookup demo tool
• JSON schemas for execution receipts and verification reports
• Marketplace-ready README and documentation

Default demo

The default demo requires no LLM API key and no network access.

It generates:

• outputs/evidence_bundle.json
• outputs/execution_receipt.json
• outputs/verification_report.json
• outputs/demo_public_key.pem

The demo public key is public. No private key is written.

Validation status

Release candidate was validated with:

• Python 3.13 full test suite: 42 passed
• module demo: valid
• independent validator CLI: valid
• CrewAI CLI crewai run: valid

Python support

This project requires:

• Python >=3.10,<3.14

Python 3.14 is not supported because CrewAI does not currently support this project on Python 3.14.

Scope and limitations

This template provides verifiable execution evidence for a tool invocation.

It does not prove semantic correctness of the tool output.
It does not prove that the policy itself is correct.
It does not protect against a compromised signer.
It does not replace sandboxing, IAM, access control, monitoring, or human approval.
It does not require or expose raw chain-of-thought.
It is not a full FDO, Gaia-X, IDS, or EDC implementation.