Tools / Decisions

open_decision()

Key

Creates a decision, pins its as_of, assembles and pins the context pack, and sets the risk level (the server may raise it by policy). From this moment, everything the decision sees is reconstructible — the pack is frozen, not re-fetched.

open_decision(question, risk_level, as_of)→ DECISION + PINNED PACK

Callable over MCP as open_decision(question, risk_level, as_of). REST mirror POST /v1/decisions Planned — alpha is MCP-only

Why your agent calls it

The moment a real question lands, the agent opens a decision — and the substrate pins everything: the as_of, the context pack, the risk level. From here on, what the decision knew is frozen and reconstructible. Decisions made outside this frame can't be audited; decisions made inside it can't be retconned.

What it returns

open_decision() · Open a decision and pin what is knowable now.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.open_decision(
    question="Should we cut the discount tier or raise onboarding spend?",
    risk_level="medium",
    as_of="2026-06-11T17:00:00Z",
)

Parameters

Param	Type	Description
questionREQUIRED	string	The decision question, plainly stated.
risk_level	enum	low · medium · high. Server may raise it by policy.
as_of	timestamp	Pin knowledge to this moment. Default now.

Example response

open_decision()JSON · ILLUSTRATIVE

{
  "decision_id": "dec_0142",
  "question": "Cut the discount tier or raise onboarding spend?",
  "as_of": "2026-06-11T17:00:00Z",
  "risk_level": "medium",
  "context_pack_id": "pack_0871",
  "status": "open"
}

Related — Decisions

keyrecord_recommendation()The full recommendation — expectation frozen.scopedrecord_decision()The immutable commit. Approval enforced server-side.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.