capx-ai/casa/casa-strategy

casa-strategy

The anchor. One durable document that every other skill reads for grounding, so the build map, the next action, and the copy all pull in the same direction. Sharp questions, with pushback on weak answers.

Steps

1. If company-brain/profile.json does not exist, tell the founder to run /casa-start

first and stop.

1. Read what exists. Load company-brain/STRATEGY.md if present, plus

profile.json, the latest decisions/, and build-map.json. If STRATEGY.md exists, you are updating it, not starting over.

1. Elicit or confirm, one sharp question at a time. Push back on vague answers.

• Target problem: the specific pain, for a specific customer, stated in their

terms. Not a market, a problem.

• Customer: who has this pain most acutely and will pay first (the beachhead).
• Approach and wedge: the narrow first thing that earns the right to the rest. Why

this wedge, why now.

• Key metrics: the one or two numbers that prove this is working (and the guardrail

that would prove it is not).

• Tracks: the two to four current lines of work, each tied to a metric.
• On-chain track (optional): if the company has or plans a token or on-chain layer,

note the traction gate that would make it eligible to tokenize.

1. Write STRATEGY.md. Save a tight document to company-brain/STRATEGY.md with those

sections, dated. Keep it short enough to reread every session.

1. Reconcile. If the new strategy contradicts a locked decision or the current build

map, name the conflict and either update the decision (record it in decisions/) or flag it for the founder. Do not silently diverge.

1. Record. Note the strategy set or revised in ledger/, dated.

Rules

• Strategy is the founder's call. You sharpen and pressure-test it; you do not invent

a direction they did not choose.

• One page. A strategy no one rereads is not an anchor.
• Every metric has a guardrail. A number that can only go up is not a strategy metric.
• No em-dashes, no emojis. Institutional tone.