Skip to content

workmagicio/workmagic

v1.0.0MIT

Ask questions about your WorkMagic attribution data, diagnose swings, compare attribution models, read lift test results, and plan media budgets — using your own WorkMagic account through the WorkMagic MCP server.

attribution-anomaly-diagnosis

Diagnose "why" questions about attribution anomalies — attribution=0 (no attributed orders), and sudden drops, spikes or retroactive changes in attributed numbers. Do NOT use for "show me numbers" — that's attribution-data-query.

attribution-custom-dimension

Handle attribution queries sliced by a business label ("by audience" / "by region" / "by brand") that is not a native field — it lives in campaign names and only becomes queryable once a Naming Convention (NC) rule is configured.

attribution-data-query

Translate explicit data requests into Cube.dev SQL and return results. Default skill for query-type asks in the Attribution domain. No estimation, no fabrication, no proactive dashboards.

attribution-edge-routing

Recognize when a user request falls outside attribution's capability boundary and route it gracefully — to another WM product (MBO / Lift Test / Creative Magic / Ads Magic / Audience Magic), to a human (CSM / DS / Eng), or to a clean decline that bridges back to what attribution can answer.

attribution-intent-clarification

Resolve genuinely ambiguous attribution queries by asking ONE focused clarifying question, then hand off to attribution-data-query (or another skill). Never become a wall of questions.

attribution-model-comparison

Pull attribution numbers under multiple attribution models side-by-side and explain why they differ in business language. Default 4-way comparison (idda / dda / last_click / platform_reported).

lift-test-readout

Read and act on completed lift test results — interpret iROAS, confidence scores, and turn the data into a post-test decision.

mbo-create-scenario

Turn a natural-language ask about media budget optimization into a configured MBO scenario — "how should I split my budget", "how much should I spend on X". Also handles rename, edit and delete on existing scenarios.

mbo-read-scenario

Interpret existing MBO scenario results — read one, compare two, or reconcile MBO against attribution, lift tests, or actual spend. Read-only; "how much should I spend" goes to mbo-create-scenario.