What problem does it solve? When running ce:work, every task gets the same one-size-fits-all treatment regardless of whether it is a new feature, a bug fix, a refactor, or an exploration. This Skill inserts an intent classification gate so each task type follows the execution strategy that actually fits it. ## Core Features & Use Cases - Automatic Intent Classification: Detects six intent types (implement, fix, refactor, explore, configure, mixed) from keywords in the plan or prompt. - Confidence-Based Gating: Applies the strategy directly at high confidence, asks a single confirmation question at medium confidence, and explicitly queries the user at low confidence. - Strategy Routing: Loads the right companion skill per intent, such as test-driven-development for new features or systematic-debugging for bug fixes, and logs the decision to state.md. - Use Case: You run ce:work with a prompt like "fix the login error". The gate classifies it as fix with 0.85 confidence, loads systematic-debugging, and enforces root-cause analysis plus a reproduction test before any code change. ## Quick Start Before running ce:work, load the intent-gate skill so it classifies my task intent and applies the matching execution strategy.
