Whitepapers

Core Protocol

• FDAA: File-Driven Agent Architecture
  The foundational spec for deterministic, auditable agent execution.

• RunProof Protocol v0.1
  Cryptographic proof generation for agent runs.

Memory & Context

• GAM: Git-Native Agent Memory v3
  Persistent, verifiable memory for AI agents using Git primitives.

Capability Control

• ACC: Agent Capability Control v2
  Declarative capability policies for agent sandboxing.

• DCT: Delegated Capability Tokens
  Time-bounded, revocable capability delegation.

Runtime Integration

• RIL: Runtime Integration Layer
  Connecting governance to agent execution runtimes.

Skill Verification

• Skill Verification v2
  Cryptographic signing and verification for agent skills.