tamper-evidence

Transform ideas into visual learning with AI-powered educational videos. Create stunning Manim animations from simple prompts for algorithms, math concepts, and more.

Privacy layer for Hermes Agent — encrypts secrets before they reach MEMORY.md, resolves them locally so plaintext never hits the LLM provider

Standalone open-source verifier for MBX v2 — AiBenchLab's tamper-evident benchmark export format. Three dependencies, zero network access, reproduces the SHA-256 content hash to confirm an .mbx.json file hasn't been altered since export.

Measuring tamper-evidence for AI agent audit logs. Hash chaining detects 0 of 5 tampering threats against an attacker who recomputes the hashes after editing; an external witness closes all five, at 0.32 ms per record.

Framework-native adapters for CCS (Correctover Conformance Shape) cryptographically signed runtime receipts — Ed25519 + JCS signed audit evidence for AI agent tool calls (Pydantic AI, CrewAI).

Tamper-evidence toolkit (npm verifyhash / verify-vh) + a zero-dependency EN 16931 / XRechnung / Factur-X e-invoice validator. Offline, permissionless, CI-gateable.

Browser demo: an agent's context window sealed with a SHA-256 hash chain, role-separated HMAC-SHA-256, and Ed25519 tool attestations, then attacked through tool output, runtime errors, and persisted state — crypto proves nothing was altered and who sent it, never whether to obey it.