agentsclimarketplace

To build brief

Skill sparkline-ventures/product-eval/skills/to-build-brief

Transform an approved decision into an agent-ready build spec, --code for Claude Code/Codex, --design for Figma/wireframe/Gemini. Use when the user says "turn this into a spec", "make it buildable", "build brief", "hand this to Claude Code / Figma", or after a brief is approved. Strips narrative voice; outputs structured, testable markdown.From its SKILL.md

Install
npx -y skills add sparkline-ventures/product-eval --skill to-build-brief

Assembled from the repository path, not quoted from the project. Check it against their README if it does not work.

One thing to look at

  • 15 stars15 stars. Stars are a popularity signal and not a quality one, but at this level it is likely that nobody has read this closely except its author, and you would be relying on your own review.

SKILL.md

2.6 KB, as published. Nobody here has run it

To Build Brief

Turn the approved Opportunity Brief / decision into a structured spec an implementation or design agent can execute. This artifact is for builders: literal requirements, screens, states, data, acceptance criteria, and metrics. Reads .product-eval/<scope>/; writes to .product-eval/<scope>/build/. Two modes; strip persuasion and prose, builders want literal, testable detail (see references/build-brief.md).

Preconditions

Only run on an approved decision. Verify a passing decision-readiness verdict exists for this item in .product-eval/<scope>/decisions-log.md; if it has not passed, refuse, do not generate a build spec for an unvalidated bet (say what is missing and point to decision-readiness / design-test). Run the conformance preflight in DATA-CONTRACT.md (Validation) on the scope files first, so a build spec is never generated from a malformed decisions-log.md or scores.md.

--code (Claude Code / Codex)

Emit: scope in/out (explicit non-goals); primary user flows; P0/P1 functional requirements each with Given/When/Then acceptance criteria; data model / key entities; states and edge/error cases; constraints, dependencies, interfaces; success metrics + instrumentation; and a "build this MVP slice first" callout. Markdown, hierarchical, no narrative voice.

--design (Figma / wireframe / Gemini)

Emit: a screen-by-screen breakdown, each view, its elements, and its states (empty / loading / error / success); flows as numbered steps; the primary action/CTA per screen; key copy; brand/platform/responsive constraints. Optionally push to Figma via its MCP path.

Workflow

  1. Load the approved brief, decision, and scores from the scope.
  2. Render the requested mode from references/build-brief.md.
  3. Carry the success metrics over, so the builder produces something measurable.
  4. Write to .product-eval/<scope>/build/ and note it's ready to hand to the target tool.

Output

The build brief in the requested mode, in plain markdown so Claude Code, Codex, Gemini, and Figma's tooling all parse it. Lead with the scope and the MVP slice. End with Next move: and name the handoff target: Codex/Claude Code for implementation, or the design tool/agent for wireframes and interaction work.

Keep looking

Skills are one crate of 627,801. Ordering is by how many stacks a row turns up in, so the top of any crate is what has actually been picked rather than what has the most stars.