What prompt helps an AI coding agent clarify a specification?
A useful specification prompt asks the assistant to expose uncertainty before it writes code. Include the user problem, intended behavior, constraints, and examples of success and failure. Ask for questions and testable acceptance criteria; do not ask it to invent business requirements to fill gaps.
What does the source say?
GitHub describes specification-driven workflows that use Markdown artifacts to guide coding agents. The reusable prompt below is an original practice exercise, not a measured popular prompt or a vendor-approved template.
How can you apply this?
Try: “Act as a critical product reviewer. Using only this brief, list missing decisions, conflicting requirements, and assumptions. Separate facts from suggestions. Propose observable acceptance examples, including one failure case. Do not implement or silently decide an unresolved requirement. Brief: [insert approved context].” Review the questions with the decision owner before asking an agent to build.
What should you watch out for?
A convincing specification is still a hypothesis about the right behavior. Keep confidential context out of tools that are not approved for it, and check every inferred requirement.