Enhance scans your project, detects your stack, and rewrites vague prompts into structured AI instructions — before you hit send.
Permission error? — no install needed.
Works with Claude Code·OpenCode·Codex CLI
5 words in. Fully structured prompt out.
|
Four steps. Zero configuration required.
Reads package.json, tsconfig, folder structure. Detects Next.js, Tailwind, Prisma, shadcn/ui — all automatically.
Parses your raw prompt to extract action, entity, and feature domain. No extra instructions needed.
Injects stack-specific rules, security practices, conventions, and relevant code context.
Review the enhanced prompt. Say no to refine further, yes to execute. You stay in control.
Years of prompt engineering best practices packed into a single slash command. No setup. No config files.
# Install globallynpm install -g @0xdevabir/enhance# Install /enhance into your AI toolsenhance setup# Inside Claude Code / OpenCode / Codex:/enhance build a login page
Enhance — SetupDetected: Claude Code, OpenCode✓ Claude Code → ~/.claude/commands/enhance.md✓ OpenCode → ~/.opencode/commands/enhance.mdReady. In any project, type:/enhance build a login page
Install in 30 seconds. Works with the AI tools you already use.