Analyze prompts for Claude Code and Codex with stage-aware coaching and rewrite suggestions.
npx skills add https://github.com/chengzhongwei/Prompt-sensei --skill prompt-sensei
Please help me install this Agent Skill. Skill: prompt-sensei Source: https://github.com/chengzhongwei/Prompt-sensei/tree/main Command: npx skills add https://github.com/chengzhongwei/Prompt-sensei --skill prompt-sensei
Prompt Sensei provides stage-aware coaching for engineers prompting AI coding agents like Claude Code and Codex, helping you craft clearer prompts, rewrite rough prompts into stronger ones, and look back on local history with privacy-preserving lookback reports.
Install in Claude Code or Codex environments, then start live coaching with /prompt-sensei observe.
High-intent search queries and answers about installing and using this skill.
You can rewrite and improve AI coding prompts by using the structured command surface to request rewrite suggestions. Prompt Sensei analyzes your inputs and provides stage-aware coaching to refine rough prompts into clearer, safer instructions for Claude Code and Codex.
Prompt coaching supports both Claude Code and Codex environments through host-integrated hooks and auto-start capabilities. It provides stage-aware feedback and rewrite suggestions tailored to the specific coding agent you are actively using.
You can review local prompt history privately by generating lookback reports that store compact coaching data locally. The skill enforces consent scopes and saves analysis results without sending your raw prompts to any external servers.
Stage-aware coaching analyzes your prompts relative to the current development stage and provides real-time feedback to improve clarity. It offers targeted rewrite suggestions to transform rough instructions into stronger prompts for Claude Code and Codex.
You can generate local, shareable coaching reports from your lookback analysis to facilitate team review. These reports are derived from compact coaching data stored locally, allowing collaborative evaluation without exposing raw prompt history.
Configuring auto-start hooks enables seamless lifecycle integration with your host tools, automatically initiating prompt analysis without manual commands. This ensures you receive consistent, stage-aware coaching feedback throughout your entire coding session.
yellow-seedPort Dockerfile RUN and ENV steps into portable shell scripts for Codex environments.
munyoung-ovoEvaluate French writing against CECRL standards with error pattern analysis.
samcdavidRestrict file edits to a designated directory using a JSON config and script hook.
guccangOptimize language model programs with declarative task definitions and automated prompt improvement.
