clarify-spec

The skill detects vague instructions automatically and prompts 2-4 clarifying questions before writing code. It analyzes project context and relevant files to narrow the scope and define the task precisely.

A structured JSON payload for prompt engineering defines a clarified task using fields like goal, scope, success criteria, and constraints. It enables deterministic prompt-architect prompts for downstream agents.

Stop repeated back-and-forth by initiating clarifying questions before code execution. Analyzing context from project files like CLAUDE.md ensures all requirements are gathered upfront for precise outcomes.

Yes, you can skip clarifications using escape hatches like phrases stating to just do it or decide independently. This proceeds with best-guess execution based on the initial instructions.

Yes, context gathering reads project files and configuration files like CLAUDE.md and AGENTS.md. This gathered context informs the targeted questions to accurately narrow the task scope.

Automated request clarification relies on available project context to formulate questions. If context is missing, questions may not fully capture the scope, requiring manual input or using escape hatches for best-guess execution.