InterAI Risk Oracle

Measured conclusion: Deep-tested. L2 outcome: tools_ok.

Pre-execution policy gate for consequential agent actions with durable trust receipts.

Add to your agent

Copy this into your MCP client config:

{
  "mcpServers": {
    "InterAI Risk Oracle": {
      "url": "https://api.interailabs.dev/mcp",
      "transport": "streamable-http"
    }
  }
}

Measured tools (9)

• oracle.verify_autonomous_action

  Primary tool for one concrete proposed autonomous action before it executes. This authenticated prepaid verification may consume account balance and records a signed trust receipt; InterAI evaluates the proposal but does not execute or modify the proposed external action. Reuse idempotency_key only when retrying the same logical request to avoid duplicate billing. Returns allow, review_required, or block with policy signals. Use oracle.verify_batch for multiple independent items and oracle.verify_response only for legacy prompt/response compatibility.

  Parameters:

  • use_case (required) — string — Stable identifier describing why the exact autonomous action is being proposed. Required together with action for the autonomous_execution contract.
  • action (required) — object — The proposed action. For executable pilot authorization use interai-canonical-action/v1 with host-registry tool_id, type, operation, exact arguments, and explicit side-effect semantics.
  • mode — string — Verification mode. Use fast_heuristic for the default low-latency path or semantic_judge when a semantic model judgment is explicitly required.
  • domain — string — Optional domain label used to provide verification context. It does not replace the exact action, context, or policy inputs.
  • policy — object — Optional strictly validated caller-scoped restrictions. They may tighten the boundary but cannot weaken host or account policy.
  • context — object — Optional runtime context surrounding the proposed action, such as environment, agent identity, counterparty, user confirmation, or other facts relevant to this decision.
  • idempotency_key — string — Stable caller-provided key for retrying the same billed verification without duplicating the economic operation. Reuse it only for the same logical request.
  • execution_context — object — Host-attested execution context bound to a canonical action. InterAI authenticates the account, not the truth of host-provided workspace, actor, run, or environment values.
  • external_evidence — array — Optional signed or verifiable external assertions. Current Stage 2 evidence is parsed, verified, bound, and recorded as non-authoritative evidence; it does not directly determine the InterAI execution decision or risk score.
  • authorization_ttl_seconds — integer — Lifetime of a single-use execution authorization for a canonical action. Only Bearer-authenticated pilot requests receive executable authorization material.

  Hints: read-only no · destructive no · idempotent no · open-world no

• oracle.verify_response

  Legacy compatibility entry point for existing clients. This authenticated prepaid verification may consume account balance and records a trust receipt, but it does not execute any proposed external action. Supply exactly one contract: prompt + response, or use_case + action; never mix them. Reuse idempotency_key only for the same logical retry. For new action preflight use oracle.verify_autonomous_action; use oracle.verify_batch for multiple items.

  Parameters:

  • mode — string — Verification mode. Use fast_heuristic for the default low-latency path or semantic_judge when a semantic model judgment is explicitly required.
  • action — object — The proposed action. For executable pilot authorization use interai-canonical-action/v1 with host-registry tool_id, type, operation, exact arguments, and explicit side-effect semantics.
  • domain — string — Optional domain label used to provide verification context. It does not replace the exact action, context, or policy inputs.
  • policy — object — Optional strictly validated caller-scoped restrictions. They may tighten the boundary but cannot weaken host or account policy.
  • prompt — string — Legacy compatibility input: the original prompt whose response is being verified. Use together with response, not with autonomous action fields.
  • context — object — Optional runtime context surrounding the proposed action, such as environment, agent identity, counterparty, user confirmation, or other facts relevant to this decision.
  • response — string — Legacy compatibility input: the response to evaluate against prompt. Use together with prompt, not with autonomous action fields.
  • use_case — string — Stable identifier describing why the exact autonomous action is being proposed. Required together with action for the autonomous_execution contract.
  • idempotency_key — string — Stable caller-provided key for retrying the same billed verification without duplicating the economic operation. Reuse it only for the same logical request.
  • execution_context — object — Host-attested execution context bound to a canonical action. InterAI authenticates the account, not the truth of host-provided workspace, actor, run, or environment values.
  • external_evidence — array — Optional signed or verifiable external assertions. Current Stage 2 evidence is parsed, verified, bound, and recorded as non-authoritative evidence; it does not directly determine the InterAI execution decision or risk score.
  • authorization_ttl_seconds — integer — Lifetime of a single-use execution authorization for a canonical action. Only Bearer-authenticated pilot requests receive executable authorization material.

  Hints: read-only no · destructive no · idempotent no · open-world no

• oracle.verify_batch

  Use for multiple independent legacy or autonomous verification items in one authenticated prepaid call (1 to 100). The batch may consume account balance and records one trust receipt per item; InterAI does not execute any proposed external action. Reuse idempotency_key only for the same logical batch retry. For one proposed action, use oracle.verify_autonomous_action instead.

  Parameters:

  • items (required) — array — Independent verification items. Each item must contain either prompt + response or use_case + action; 1 to 100 items are accepted.
  • idempotency_key — string — Stable caller-provided key for retrying the same billed verification without duplicating the economic operation. Reuse it only for the same logical request.

  Hints: read-only no · destructive no · idempotent no · open-world no

• oracle.get_pricing read-only

  Read-only public metadata for costs, trial availability, idempotency, and top-up or payment paths. Use it before budgeting or onboarding; it does not evaluate risk or create a trust receipt.

  Hints: read-only yes · destructive no · idempotent yes · open-world no

• oracle.discovery_bundle read-only

  Read-only bootstrap or handoff bundle with service identity, interfaces, runtime mode, contracts, schemas, and sample payloads in one response. Use it for complete integration context; use oracle.service_descriptor for focused service metadata, oracle.agent_card for A2A capabilities, or oracle.get_pricing for costs.

  Hints: read-only yes · destructive no · idempotent yes · open-world no

• oracle.service_descriptor read-only

  Read-only general service descriptor with identity, endpoints, billing, trust, and supported agent protocols. Use it to inspect service-level integration metadata; use oracle.agent_card for A2A skills or oracle.discovery_bundle for the complete bootstrap package.

  Hints: read-only yes · destructive no · idempotent yes · open-world no

• oracle.agent_card read-only

  Read-only A2A Agent Card with agent identity, security scheme, capabilities, and skills. Use it when selecting or integrating InterAI through A2A; use oracle.service_descriptor for general service metadata or oracle.discovery_bundle for complete bootstrap context.

  Hints: read-only yes · destructive no · idempotent yes · open-world no

• oracle.get_trust_receipt read-only

  Read-only lookup of the canonical public representation and signature metadata for a trust receipt that already exists. Supply receipt_id. It does not create or re-run a decision; use oracle.verify_trust_receipt_signature when the goal is signature validation.

  Parameters:

  • receipt_id (required) — string — Identifier of the existing InterAI trust receipt whose canonical public representation and signature metadata should be returned.

  Hints: read-only yes · destructive no · idempotent yes · open-world no

• oracle.verify_trust_receipt_signature read-only

  Read-only service-side validation of a trust receipt that already exists. Supply signature plus either receipt_id for the canonical server-stored receipt or a caller-supplied receipt object. Returns valid true or false when signing is enabled. Current signatures use HMAC-SHA256 and are service-verifiable by InterAI, not independent offline public-key proofs. It does not evaluate risk or create a receipt; use oracle.get_trust_receipt to retrieve receipt metadata.

  Parameters:

  • signature (required) — string — HMAC-SHA256 signature to validate against the referenced InterAI receipt. Required.
  • receipt — object — Optional caller-supplied receipt object, typically evidence received from another agent or system. Either receipt or receipt_id is required.
  • receipt_id — string — Optional ID of an existing InterAI trust receipt. InterAI resolves the canonical server-stored record before validating the supplied signature. Either receipt_id or receipt is required.
  • signature_alg — string — Optional signature algorithm declaration. The current supported value is hmac-sha256.
  • signed_payload — string — Optional exact serialized payload that was signed. Supply it when validating externally transported receipt evidence that includes the original signed payload.

  Hints: read-only yes · destructive no · idempotent yes · open-world no

Repository

Measured by MCPYP probing. Untested fields are omitted, never estimated. Listing is always free; MCPYP never proxies calls or stores user tokens.