What problem does it solve? Teams often jump into implementation before fully understanding a problem, leading to rework and misaligned designs. This Skill provides a structured exploration mode where the AI acts as a thinking partner—investigating the codebase, visualizing options with diagrams, and clarifying requirements before any code is written. ## Core Features & Use Cases - Exploration stance: Asks open questions, challenges assumptions, compares options with tradeoff tables, and draws ASCII diagrams of architectures, state machines, and data flows. - OpenSpec integration: Reads project context and rules from openspec/config.yaml, checks active changes via openspec list --json, and captures crystallized decisions into proposals, designs, specs, and tasks using the openspec CLI. - Guardrails against premature implementation: Never writes application code in explore mode; it only investigates and optionally scaffolds change artifacts through openspec new change. - Use Case: A developer is unsure whether to use SQLite or Postgres for a CLI tool. The Skill maps constraints, builds a comparison table, recommends a path, and offers to capture the decision in a design artifact. ## Quick Start Ask the agent to enter explore mode and think through whether your planned feature should use approach A or B before creating a change proposal.