SuperAPI AI Directory

starlight-ceo-review

Agents & models

View repository

Updated 10-16 05:47

About this skill

Starlight CEO Review

Purpose

Test whether a proposal deserves to exist and whether its scope serves the intended user without transferring hidden cost, liability, or support burden to the operator.

When it fires

  • A material feature, product, partnership, or platform investment needs a decision.
  • The plan contains unverified market, cost, or adoption assumptions.
  • Scope is either artificially narrow or expanding without evidence.

Inputs

  • Target user, job to be done, and desired behavior change.
  • Evidence of demand and alternatives already available.
  • Revenue or mission hypothesis, operating cost, support owner, and constraints.
  • Reversibility, legal/privacy implications, and success thresholds.

Workflow

  1. State the decision and the evidence required to make it.
  2. Test the user problem and the proposed mechanism separately.
  3. Compare build, integrate, partner, defer, and do-nothing options.
  4. Identify the smallest coherent scope that delivers the full promised outcome.
  5. Surface lifecycle cost: maintenance, support, security, compliance, and migration.
  6. Define leading indicators, stop conditions, and the next human decision.

Output contract

Return a verdict of build, revise, defer, or stop, followed by evidence, assumptions, strategic fit, alternatives, scope corrections, operating-owner obligations, metrics, and human gates.

Tools & MCP

Use current product evidence, customer research, financial assumptions, and primary market sources. Do not invent demand, ROI, partner approval, or “zero liability.”

Quality bar

  • Ambition is connected to a concrete user outcome.
  • Completeness means the promise works end to end, not that every imaginable feature ships.
  • Human and agent effort estimates include uncertainty and operating cost.
  • Irreversible or public decisions remain human-owned.

Example

Input: “Should we fork an agent runtime?”

Good output: compare a profile/adapter layer against a fork, price the maintenance and support obligations, define the evidence that would justify owning the runtime, and recommend the reversible path until that evidence exists.


Built on SIP — Starlight Intelligence Protocol Vertical: starlight-agent-skills · portable capability layer