Skills Governance

Skills governance controls which declared agent skills can be invoked. It applies Arsenal-style policy to SKILLS.md files and runtime skill calls.

What it gates

Dimension	What it controls
allow / deny	Whether the skill can be invoked at all
rate limit	Maximum invocations per sliding window
argument constraints	JSON Schema validation of invocation arguments
time windows	UTC hours and weekdays when invocation is allowed
consent	Human approval requirements for high-risk skills
audit level	No audit, hash-only audit, or full payload audit

Policy example

version: 1
agent: did:oas:l1fe:agent:frontend-bot
skills:
  frontend-design:
    allow: true
    rate_limit:
      window: 1h
      max: 10
    audit_level: hash
  deploy-production:
    allow: true
    require_consent: true
    audit_level: full
default:
  allow: false

Runtime use

const skills = agent.skillsPolicy();

if (skills.canInvoke('frontend-design')) {
  await runFrontendDesign();
}

skills.assertCanInvoke('deploy-production');

Audit receipts

Recorded invocations append immutable receipts to a BLAKE3 hash chain. hash audit mode records safe metadata and a payload hash. full audit mode records the argument payload for high-sensitivity forensic workflows.

Relationship to Arsenal

Skills governance is not a separate security primitive. It is an Arsenal policy specialization for “which skill may this agent invoke right now?”

Wallet Derivation Conformance

Getting Started

Feature Stack

OpenAgent SDK

OpenAgent Client

OpenAgent Server

What it gates

Policy example

Runtime use

Audit receipts

Relationship to Arsenal

​What it gates

​Policy example

​Runtime use

​Audit receipts

​Relationship to Arsenal

What it gates

Policy example

Runtime use

Audit receipts

Relationship to Arsenal