Works with the AI Agents you already use
Claude Code
Cursor
Codex CLI
Muse
Dots+21 moreClaude CodeCursorCodex CLIMuseDots
Manus
Grok
Grok Bot+18 agentsIdea → Requirements Spec (PRD)
Turns a raw app idea into an approved idea brief and a requirements spec with stable FR IDs, testable acceptance criteria and an explicit out-of-scope list.
Free
Idea → Requirements Spec (PRD)
Example session with this skill installed
I want an app that stops my household from wasting groceries. We forget what's in the fridge and things go off. Two of us share the shopping. No recipes, I just want to know what to use first. Let's spec it — call it ShelfLife.
- Read your context and instructions
- Compiled the idea requirements spec
- Generated the document
One-sentence test first
ShelfLife is a pantry tracker that warns a household before food expires.
That passes, and you've already drawn the most important line (no recipes). Two things I'd push on before writing anything:
Riskiest assumption: that people will log items. If adding an item takes more than a few seconds, nobody will do it after week one. I'm treating "name + rough expiry, nothing else" as a hard constraint.
"Shared" needs a definition. Two people seeing one list means some notion of a household and a way to join it. I'll keep it minimal and list the details as open questions.
Here's the brief
Idea — ShelfLife
One-Sentence Summary
ShelfLife is a pantry tracker that warns a household before food expires.
Problem
Two-adult households throw away food they forgot they had. Today they rely on memory and the fridge door; expiry dates are printed on packaging nobody reads until it's too late. The cost is money and guilt, a few times a week.
Vision
Adding an item takes seconds: name, rough expiry, done. ShelfLife sorts everything by what's going off next and sends one calm daily nudge, never a stream of alerts. Both people in the household see the same list. It does one job: make "what do we need to use up?" answerable in five seconds.
Target User
- Primary: a two-adult household that shops weekly
- Secondary: a single person meal-prepping
Key Features (high-level)
- Quick add with an estimated expiry
- "Use soon" list sorted by expiry
- One daily digest notification
- A shared household list
Non-Goals
- Recipes or meal planning
- Barcode scanning (v2 candidate)
- Shopping lists
Open Questions
- Do items have quantities, or just presence?
- How does a second person join a household?
Saved to program/shelflife/specs/idea.md.
Next: approve this brief (or tell me what to change), and answer the two open questions if you have a view. Then I'll write requirements.md with FR-numbered user stories and testable acceptance criteria.
idea-requirements-spec-prd.pdf
PDF · document
Example file from a real run - the skill writes it into your workspace.
Connects securely to your tools. The creator never sees your data.
About this skill
Most product ideas stall, or turn into sprawling AI-generated PRDs nobody can build, because nobody pinned down who it's for, what problem it solves, and what's out of scope. This skill makes your agent act like a sharp product manager: it pressure-tests the idea first and only then writes two Markdown documents a team or a coding agent can build from.
What it does
- Runs the one-sentence test. No document gets written until the product can be described in one sentence, with a real target user and a real problem.
- Plays devil's advocate on the weakest assumption and proposes a v1 scope with an explicit not-now list.
- Writes
specs/idea.md: summary, problem, vision, target users, key features, non-goals, open questions. - Writes
specs/requirements.md:FR-nuser stories, observable acceptance criteria (including rejection paths), non-functional targets, constraints, assumptions, out-of-scope and a glossary. - Keeps IDs stable so later plans, tests and audits can cite
FR-3and mean the same thing forever.
How it works
- Discovery. You describe the idea; the agent asks a few short questions (who has the problem, what they do today, what's in v1).
- Idea brief. Once the one-sentence test passes, it writes
idea.md. - Approval gate. You approve before anything else happens. Silence is not approval.
- Requirements. It resolves the open questions (asking only about the ones that change product behaviour) and writes
requirements.md.
Frameworks & tools
Tech-agnostic on purpose. Markdown only: no code, no stack choices, no scripts, no network access.
Why this beats prompting it yourself
"Write me a PRD" gets you a confident wishlist. This skill refuses to write until the idea is sharp, forces a scope boundary, requires every acceptance criterion to be observable (with at least one rejection path), and strips stack words out of requirements so they don't lock you into a design before you've made one.
Use cases
- Turn a shower thought into a PRD before you open an editor.
- Scope a lean MVP and write down what you're not building.
- Give a coding agent a spec it can't misread.
- Create a backlog of stable, testable user stories.
Known limitations
Doesn't write code, tests, UI mocks, architecture or database schemas, and won't pick a tech stack. That's deliberate. (The paid App Blueprint skill continues from here through design, architecture and schema.)
How to use
- 1
Connect your agent to the Agensi MCP
Copy the prompt and paste it into Claude, ChatGPT, Cursor or any agent that supports MCP. It sets up the connection for you. Manual setup
- 2
Sign in to Agensi
Your agent shows a sign-in link. Sign in and click Allow.
- 3
Ask for what you need
Your agent loads this skill and uses it on your task. It always asks before any payment.
Reviews
No reviews yet
Be one of the first to try it. Every listed skill goes through the trust checks below.
Security scanned
Checked by our 8-point scan before listing
1 use
Times used to date
Free forever
No account required to browse
Trust & safety
Security scan
Security scan passed 3 days ago
- Free with an Agensi account