What problem does it solve? Ambiguous, conflicting, or underspecified development requests cause agents to guess, expand scope, or build the wrong thing. This Skill turns a raw request into an evidence-grounded understanding with explicit assumptions, resolved conflicts, and a deterministic proceed-or-block decision before any code changes happen. ## Core Features & Use Cases - Evidence-first clarification: Inspects repository instructions, configuration, tests, and nearby implementation before asking the user anything, so questions never cover facts already discoverable in the project. - Bounded question rounds: Enforces configurable limits (maxQuestionsPerRound, questionsPerPrompt, askEveryPrompt) so clarification stays concise and never exhausts its budget on unnecessary questions. - Plan confirmation gate: When confirmPlanBeforeExecution is enabled, states the objective, affected files, and risks, then waits for explicit user confirmation before downstream skills run. - Use Case: A user asks to "add authentication" to an existing API. The Skill inspects the codebase, finds conflicting session and token patterns, asks targeted questions about backward compatibility, records safe defaults for naming, and outputs a validated clarification report for the planning workflow. ## Quick Start Ask the agent to clarify my request to add caching to the API before writing any code, using the clarify-the-ask skill.
