SmartMemory
GET
/memory/decisions

Get all active decisions, optionally filtered.

With provenance_memory_id set, routes to the inverse-provenance query (CORE-DECISION-PROVENANCE-LOOKUP-1); all filtering is delegated to the core method so filter+limit composition is correct. Without it, behaviour is unchanged from the original list path.

Query Parameters

domain?|

Filter by domain

decision_type?|

Filter by decision type

min_confidence?Min Confidence

Minimum confidence

limit?Limit

Maximum results

agent_id?|

Filter by producing agent id (CORE-AGENT-ATTRIBUTION-1). NULL-valued rows are excluded when set.

provenance_memory_id?|

Inverse-provenance filter (CORE-DECISION-PROVENANCE-LOOKUP-1). When set, returns only active decisions whose provenance subgraph (within max_depth hops) contains this memory id. Combined with the other filters: domain/decision_type/min_confidence are applied in-query so the response never silently truncates below limit when more matches exist. Unknown or out-of-scope memory ids return an empty list (never 404) to avoid existence leakage.

Response Body

application/json

application/json

curl -X GET "https://example.com/memory/decisions"
null
{
  "detail": [
    {
      "loc": [
        "string"
      ],
      "msg": "string",
      "type": "string",
      "input": null,
      "ctx": {}
    }
  ]
}