Define project requirements with AI
Work through goals, behaviours and constraints in a paced conversation, saving completed sections to a local requirements document as the decisions become clear.
Define project requirements with AI when a team agrees it needs a room-booking tool but has not yet decided how competing requests should be handled. The requirement-elicitation skill guides an agent through manageable rounds of questions. It adapts the details to the kind of project rather than asking a mobile team and a command-line developer the same things.
A useful answer might change the scope. If offline booking is essential, that affects more than the wording on the submit button. The discussion can return to an earlier point, record a constraint or leave a question unresolved. Completed sections accumulate in a local file so the team can resume later.
What it does
- Identify the user, goal and boundaries of the first version.
- Describe what should happen during conflicting or incomplete requests.
- Separate functional behaviour from performance and access constraints.
- Revisit an earlier answer when a later decision changes its meaning.
- Keep unresolved questions alongside the decisions they may affect.
Install and use
- Provide a project idea, its audience and any constraints already known. The agent needs a writable project folder for the local requirements document.
- Ask, “Help define our room-booking tool. Start with first-release scope, then explore two users requesting the same room.”
- Read each section before moving on and review the assembled document with the team.
FAQ
Source
Made by Britt Crawford. Original skill: requirement-elicitation, published on GitHub under the MIT licence.
View on GitHub