Why SHA-256 Isn't Enough for AI Agent Governance (And How We Fixed It)

Outline & Key Points:

1. The Problem: Explain why self-signed SHA-256 digests fail as authentication (anyone with write access can recompute). Use the audit's finding: "integrity ≠ authentication".
2. The Threat Model: Briefly describe attack vectors (bundle tampering, rollback, unauthorized issuer).
3. The Solution: Walk through Ed25519 signing + trust store verification. Show code snippets of signing.py and runtime verification.
4. Beyond Signing: Mention ActionRequest model and unified integrations as complementary hardening measures.
5. Lessons Learned: Be honest about what took time (e.g., fixing TOCTOU, normalizing enums, aligning release discipline). This builds credibility.
6. Call to Action: Link to GitHub repo, PyPI, and invite discussion in comments.

Tags: #python, #ai, #security, #opensource, #langchain, #crewai

---