Tools / Causal

propose_edge()

Key

Proposes a cause→effect edge with rationale and evidence links. Lands at status=proposed. Nothing the system computes ever runs on an edge a human hasn't approved.

propose_edge(cause, effect, rationale, evidence_ids)→ PROPOSED EDGE

Callable over MCP as propose_edge(cause, effect, rationale, evidence_ids). REST mirror POST /v1/graph/edges Planned — alpha is MCP-only

Why your agent calls it

The agent reads a post-mortem and suspects discount drove win rate. It proposes the edge, citing the evidence. Until a human approves it, nothing downstream — no identification, no estimate, no simulation — will run on it.

What it returns

propose_edge() · Propose a cause→effect edge, with evidence.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.propose_edge(
    cause="discount",
    effect="win_rate",
    rationale="Margin data shows the tier is net-negative.",
    evidence_ids=["mem_0871"],
)

Parameters

Param	Type	Description
causeREQUIRED	string	The cause variable.
effectREQUIRED	string	The effect variable.
rationaleREQUIRED	string	The claimed mechanism.
evidence_ids	string[]	Memory or sources supporting the edge.

Example response

propose_edge()JSON · ILLUSTRATIVE

{
  "edge_id": "edg_0455",
  "edge": "discount → win_rate",
  "status": "proposed",
  "evidence": ["mem_0871"]
}

Related — Causal

keypropose_nodes()Propose variables for the graph — status proposed.keyget_graph()The governed graph, by version or as-of.keyvalidate_graph()Structural checks before anyone trusts an edge.keyidentify()Can this effect be estimated at all? The gate.keyestimate()A causal estimate that survives the skeptic checks.keysimulate_intervention()do(X=x) on the approved graph — the knock-on report.keyattribute_change()What caused this metric to move?