Skip to content

introduction.md

Latest commit

 

History

History
45 lines (32 loc) · 1.53 KB

introduction.md

File metadata and controls

45 lines (32 loc) · 1.53 KB

HushSpec

HushSpec is a portable, open specification for declaring security rules at the tool boundary of AI agent runtimes. It defines what security rules an agent operates under, without prescribing how those rules are enforced.

Why HushSpec?

AI agents interact with tools — file systems, network APIs, shell commands, MCP servers. HushSpec provides a standard way to declare which interactions are allowed, blocked, or require confirmation.

  • Portable: Works with any engine that implements the spec
  • Declarative: Rules are stateless YAML — no runtime state, no detection algorithms
  • Fail-closed: Unknown fields are rejected; invalid documents produce errors, not silent misconfiguration
  • Extensible: Optional modules for posture state machines, origin-aware profiles, and detection thresholds

A Minimal Example

hushspec: "0.1.0"
name: my-policy

rules:
  forbidden_paths:
    patterns:
      - "**/.ssh/**"
      - "**/.aws/**"

  egress:
    allow:
      - "api.openai.com"
      - "*.anthropic.com"
    default: block

  tool_access:
    block:
      - shell_exec
      - run_command
    default: allow

Spec vs Engine

HushSpec defines the portable rule language. Engines like Clawdstrike implement the spec and add engine-specific features (detection algorithms, receipt signing, async guard infrastructure).

Current Status

HushSpec v0.1.0 — unstable. Breaking changes may occur between minor versions. The spec will stabilize at v1.0.0.