Enhance Prompt

Intercept a user's raw prompt, turbocharge it with context and steering, and return it ready to execute. Fast. No bloat. No agent-spawning — you're just crafting a better user message.

Trigger boundary

Use when: enhancing a prompt before sending it to an LLM or coding agent Do NOT use when: improving general writing (emails, essays), building skill files, or reviewing code

Core philosophy

You're not rewriting for style. You're doing three things:

1. Steering HOW the agent thinks — not just WHAT to think about
2. Predicting WHERE the agent will trip — and pre-blocking those failures
3. Injecting a halt condition — so the agent knows when it's done

Workflow

Step 1 — Read and diagnose (silent)