Produces a short, structured Discovery Brief from a task description and available context. The brief clarifies the real problem to solve, separates assumptions from constraints, and proposes testable hypotheses so downstream authors and engineers can act without re-framing.
Use this skill at the start of exploratory work: when a request is ambiguous, before gathering evidence, or when multiple plausible solutions exist. Ideal for analysts, PRD authors, and early-stage product discovery to reduce rework later.
Likely to be useful in agentic workflows that orchestrate multi-stage reasoning (Claude Code, Codex-based agents, and local orchestrators). This is an internal methodology skill rather than a user-facing UI component.
Ideate is a methodology-only skill that guides an analyst agent through framing a Discovery Brief — problem, constraints, hypotheses, and open questions. No scripts to run; the entire skill is a structured markdown prompt with a concrete output template. Clean and straightforward but thin in depth.
Pure methodology skill with no executable code. Zero security surface. Well-structured frontmatter and clear output contract. Minimal but functional — does exactly what it says and nothing more.