When an agent acts,
who is liable?
Autonomous agents sign contracts, make commitments, and occasionally fail in ways that create legal exposure. Operators, framework vendors, and model providers are all pointing at each other. iAgentLaw is the proposed standards + tooling + insurance layer to resolve it. Parked pending a law-firm anchor partnership.
iAgentLaw would provide four primitives
Terms of Agent templates
Pre-drafted operator agreements, end-user terms for agent-delivered services, and agent-to-agent transaction frameworks covering EU AI Act, US state AI laws, and common international regimes.
Action risk scoring
Pre-action API: submit a proposed action + jurisdiction, receive a legal-risk score and rationale. Integrates with iAgentSafe as a runtime guardrail.
Insurance brokerage
Partnership with specialty underwriters. Agent operators can obtain E&O, cyber, and professional liability coverage priced on iAgentFi ratings + iAgentSafe posture.
Incident arbitration
Pre-negotiated dispute resolution protocol for agent-involved incidents. Evidence kit generation from iAgentLog traces.
The gap iAgentLaw fills
The liability vacuum is real
No existing framework cleanly assigns responsibility when an autonomous agent causes harm. First-mover defining the answer wins defaults.
Enterprises can't deploy without a resolution
Fortune 500 legal teams are currently blocking agent deployments because risk is unquantified. A productized legal+insurance layer would be the unlock. iAgentLaw is parked until a law-firm anchor partner makes the build credible.
Would compound with the stack
If activated, iAgentFi ratings would inform insurance pricing, iAgentLog traces would become evidence kits, iAgentSafe policies would become the compliance layer, and iAgentLaw would be the regulatory substrate. The stack integrations wait for a partnership to make this real.
Planned endpoints in Parked pending law-firm partnership
iAgentLaw is parked pending a law-firm anchor partnership. OpenAPI specification at /.well-known/openapi.yaml previews the planned API surface if a partnership materializes.
POST /v1/risk/evaluate |
Pre-action legal risk scoring |
GET /v1/templates/tos |
Terms of Agent template library |
POST /v1/insurance/quote |
Request insurance quote for agent operations |
POST /v1/incidents/file |
File an incident for arbitration |
GET /v1/jurisdictions |
Supported jurisdictions and frameworks |
iAgentLaw is one layer
Sixteen products. One stack. One entity. Trust, discovery, observability, payments, safety, simulation, composition, memory, identity, legal, markets, and owned compute underneath. Each layer reinforces the others. Use one or use them all.