Dojo

Per-run Safe House detector posture (staff-only).

Returns a run's Safe House detector posture — the WORST health across the run's canonical safe_house_decisions (tool_call ∪ tool_response), plus screened/expected and an alarm flag. The dashboard self-fetches this so a degraded / failed-open sensor never renders green (the MNE-913 fail-open-looks-healthy class; C6 AC2). Admin-only; fail-CLOSED — a read error is a 500, never a fake-clean posture. MNE-1347.

get/dojo/safe-house-posture

Request

Base URL: https://api.mnemom.ai/v1
URL: https://api.mnemom.ai/v1/dojo/safe-house-posture
Auth: HTTP bearer

Query parameters

agentstring required

The agent id (mnm-…).

runstring

Optional gateway run id (x-mnemom-session). Omit for agent-only posture (the most recent run) — the dashboard has the agent id but NOT the gateway run id (different id-space).

Response

The run's detector posture.

sourcestring required

Changes

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