memory

Load Guardrails

Deterministically load a context's guardrail set for a client-side hook.

The REST twin of the MCP load_guardrails tool and /pinned's sibling: the trusted-tier pinned set (delivery_mode='always') plus the memories marked with details.tool_trigger, each list ordered importance DESC, created_at ASC, id ASC and capped on its own (cap bounds tool_triggered only; the pinned list uses settings.pinned_load_cap). Plain SQL — no ranking, no rerank. Items are L1 (+ L2 on pinned items) and carry the normalized tool_trigger object as data; the server never runs the patterns. The response's format and version are the shared cache contract documented with the MCP tool.

Request: {"context_id": "<context-uuid>", "cap": 50}

Example: POST /api/v1/memory/guardrails Authorization: Bearer <api_key>

post/api/v1/memory/guardrails

Request

  • Base URL: https://memory.kagura-ai.com (the document declares no server; this is the origin it was published from)
  • URL: https://memory.kagura-ai.com/api/v1/memory/guardrails
  • Auth: none declared

Headers

authorizationstring nullable

Request body

context_idstring nullable

Context UUID whose guardrail set to load

capinteger nullable

Optional override for the max tool-triggered memories returned (bounded)

Response

Successful Response

statusstring
formatinteger required
versionstring required
total_availableinteger required
truncatedboolean required
capinteger required
pinned_capinteger required
pinned_total_availableinteger required
pinned_truncatedboolean required
tool_triggered_total_availableinteger required
tool_triggered_truncatedboolean required

Changes

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