What problem does it solve?
Vague, ambiguous, or incomplete prompts sent to the Codex CLI often lead to guesswork, incorrect outputs, and wasted iteration time. This skill eliminates that friction by automatically detecting low-quality prompts, inferring missing context, and generating optimized, structured prompts that produce reliable, accurate results.
Core Features & Use Cases
- Low-quality prompt detection: Auto-activates when user input has 2 or more signals of being vague, such as missing goals, undefined scope, unclear success criteria, or mixed unrelated asks.
- 4-dimension prompt analysis: Extracts and fills gaps in intent, scope, constraints, and done-state to create complete task definitions.
- Structured prompt assembly: Generates XML-formatted Codex prompts with tailored contracts, verification loops, and safety rules matched to the task type (debug, refactor, implement, review, etc.).
- Use case: If you submit the prompt "fix the auth bug" to Codex, this skill will ask one targeted clarifying question if needed, then produce a prompt with explicit scope, verification steps, and safety guardrails to deliver a correct, minimal fix without unnecessary changes.
Quick Start
Use the codex-prompt-optimize skill to transform your vague Codex request like "make the backend faster" into a clear, actionable prompt with defined scope and success criteria.
