Docs/AI Trading Agents/Skill System and SDK Surface

Skill System and SDK Surface

Production Track

Skill modules extend agent capability while preserving guardrail contracts.

Objective

Allow extensibility for ecosystem growth without introducing unbounded execution vectors.

Skill Lifecycle

Skills can be listed, installed, versioned, and revoked per agent profile.

Critical guard skills remain mandatory and cannot be removed in protected environments.

SDK Coverage

Rust SDK provides direct system-level integration, TypeScript SDK supports app and workflow tooling, and Python SDK supports research and automation workflows.

SDK methods map to proposal generation, status inspection, and controlled execution requests.

Ecosystem Integration

MCP and external orchestrator integrations should route through controlled tool contracts.

Skill packaging should include metadata on required permissions and safety class.

Production Checkpoints

  • Skill install/removal is versioned
  • Guard modules are non-removable where required
  • SDK actions map to audited APIs
  • Permission metadata enforced at runtime

Continue exploring

AI Trading Agents