Tools / Decisions

record_decision()

Scoped

The final commit. For risk ≥ medium it requires a granted approval — checked server-side from the decision row, not from tool arguments (invariant I6). Once written, the decision is immutable.

record_decision(decision_id, approval_id)→ IMMUTABLE DECISION

Callable over MCP as record_decision(decision_id, approval_id). REST mirror POST /v1/decisions/commit Planned — alpha is MCP-only

Why your agent calls it

The commit. For anything medium-risk or above, the server checks for a granted approval itself — the requirement is computed from the decision row, so an agent can't talk its way past it. After this write, the decision is immutable.

What it returns

record_decision() · The immutable commit. Approval enforced server-side.ILLUSTRATIVE

Call it from an agent

Over MCP, your agent invokes it on its own when the moment calls for it — just describe the situation. A Python SDK for headless runtimes is planned; this is what the call will look like.

agent · python sdkSDK · PLANNED — ALPHA IS MCP-ONLY

from cascadian import Cascadian

cx = Cascadian(api_key=os.environ["CASCADIAN_API_KEY"])

result = cx.record_decision(
    decision_id="dec_0142",
    approval_id="apr_0027",
)

Parameters

Param	Type	Description
decision_idREQUIRED	string	The decision to commit.
approval_id	string	Required for risk ≥ medium. Verified server-side.

Example response

record_decision()JSON · ILLUSTRATIVE

{
  "decision_id": "dec_0142",
  "status": "decided",
  "approval": "apr_0027 — verified server-side",
  "immutable": true
}

Related — Decisions

keyopen_decision()Open a decision and pin what is knowable now.keyrecord_recommendation()The full recommendation — expectation frozen.keyrecord_outcome()Measured actuals against the frozen expectation.keypropose_learning()Queue what the variance taught — humans approve.keyreview_calibration()How good are the calls, over time and by type.keyget_decision()The full decision card, by id.