Risk

Assess risk for a single agent action

Compute an individual risk assessment for an agent about to take an action. Requires JWT auth and the risk_assessment feature flag on the caller's plan. Dispatches a ZK proof request asynchronously (best-effort; check /risk/proofs/{proof_id} to retrieve). Cached for 5 minutes per {agent_id, action_type, risk_tolerance, amount} tuple. An agent whose reputation is private can be assessed only by a member of its governing org (403 reputation_private otherwise); a tombstoned agent returns 404.

post/risk/assess

Request

  • Base URL: https://api.mnemom.ai/v1
  • URL: https://api.mnemom.ai/v1/risk/assess
  • Auth: HTTP bearer

Request body

agent_idstring required
source'api' | 'playground'

Response

Risk assessment.

assessment_idstring
agent_idstring
risk_scorenumber

0–1000 integer (higher = riskier).

risk_level'low' | 'moderate' | 'high' | 'critical'
recommendation'proceed' | 'proceed_with_caveats' | 'require_approval' | 'block'
confidence'insufficient' | 'low' | 'medium' | 'high'
contributing_factorsobject[]
suggested_thresholdsobject
explanationstring
proof_idstring nullable
proof_status'none' | 'pending' | 'completed' | 'failed'
created_atstring date-time

Changes

No recorded changes to this operation across all 1 revision of this API.