---

name: scout description: Quick project reconnaissance and health check via @geepers_scout - identifies issues, quick wins, and improvement opportunities. version: 1.0.0

Project Scout

You are performing a quick reconnaissance mission on the current project. This is a lightweight health check that identifies issues, quick wins, and generates actionable improvement reports.

When to Scout

• Mid-session checkpoints (every 1-1.5 hours of active work)
• After completing a feature - verify nothing broke
• When picking up unfamiliar code - understand current state
• Before major changes - assess risk areas
• When something feels "off" - systematic diagnosis

Scout Mission

Launch @geepers_scout with the current project context.

The scout agent will:

1. Scan for Issues

   • Syntax errors, broken imports
   • Missing dependencies
   • Configuration problems
   • Path and reference issues

2. Identify Quick Wins

   • Low-hanging fruit improvements
   • Obvious fixes (<30 min effort)
   • Documentation gaps
   • Code cleanup opportunities

3. Review Recent Changes

   • Git history analysis
   • New files and modifications
   • Potential regressions

4. Generate Improvement Report

   • Prioritized recommendations
   • Effort estimates
   • Impact assessments

Output Location

Scout generates reports in:

• ~/geepers/reports/scout-{project}-{timestamp}.md
• Summary displayed in conversation

Output Format

Copy🔍 SCOUT REPORT: {project}

📊 Health Score: [A-F grade]

🔴 Critical Issues:
- [Issue requiring immediate attention]

🟡 Warnings:
- [Issues that should be addressed soon]

⚡ Quick Wins (< 30 min):
1. [Quick fix with high impact]
2. [Easy improvement]
3. [Documentation gap]

📈 Improvement Opportunities:
- [Medium-term enhancement]
- [Architecture improvement]

📁 Report saved: ~/geepers/reports/scout-{project}-{date}.md

Scout vs. Full Session Start

Use Scout for quick checks during work. Use Session Start when beginning a new session.

Key Principles

1. Fast execution - Scout should complete in under 2 minutes
2. Actionable output - Every finding has a clear next step
3. Non-disruptive - Scout observes, doesn't modify
4. Prioritized findings - Critical issues first
5. Lightweight - Minimal agent overhead

Common Scout Triggers

• "What needs work?"
• "Anything broken?"
• "Quick health check"
• "Scan for issues"
• "What's the status?"
• "Any quick wins?"
• After completing a task: "Did I break anything?"