LaunchKit · 2026
Back to Skills

reprompter

Transform messy prompts into well-structured, effective prompts — single or multi-agent.

0
470 downloads
by @aytuncyildizli

Setup & Installation

openclaw skills install @aytuncyildizli/reprompter

Or with the ClawHub CLI, for registry-managed skill folders outside a full OpenClaw workspace:

npx clawhub install reprompter

What This Skill Does

RePrompter restructures rough prompts into XML-formatted prompts using task-specific templates and quality scoring. It has two modes: single prompt improvement with before/after scoring, and Repromptception, which plans a multi-agent team, writes per-agent sub-prompts, launches them via tmux, and evaluates outputs. Scores six dimensions including clarity, specificity, and verifiability.

Structuring prompts with templates and quality scoring before execution reduces agent misinterpretation and avoids iterative back-and-forth from underspecified inputs.

When to use it

  • Turning a vague bug description into a structured bugfix prompt before sending to an agent
  • Coordinating a parallel frontend and backend audit with separate specialized agents
  • Scoring an existing prompt to identify weak specificity or missing constraints
  • Generating per-agent XML sub-prompts for a sequential pipeline task
  • Reformatting a rough feature idea into a role-context-requirements template

Example Workflow

Here's how your AI assistant might use this skill in practice.

INPUT

User asks: reprompt this: fix the login bug on /auth route

AGENT
  1. 1Detects task type as bugfix and selects bugfix-template.md
  2. 2Runs Smart Interview to ask about execution mode and output format
  3. 3Auto-detects tech stack from package.json in the current working directory
  4. 4Generates XML-structured prompt with role, context, symptoms, requirements, and success criteria
  5. 5Displays before/after quality score table and offers to execute or copy
OUTPUT

Structured XML bugfix prompt scoring 8.5/10 versus the original 2/10, ready to send to an agent or copy

Requirements

Accounts, API keys, or tools you or your AI assistant may need to set up while using this skill.

tmux installed (required for Repromptception multi-agent mode only)CLAUDE_CODE_EXPERIMENTAL_AGENT_TEAMS=1 set in Claude Code settings (required for Repromptception multi-agent mode only)