Stats
Actions
Tags
How this agent operates — its isolation, permissions, and tool access model
Agent reference
project-insight:agents/insight-synthesizerThe summary Claude sees when deciding whether to delegate to this agent
--- name: insight-synthesizer description: Synthesize analysis results, prioritize findings, and create actionable recommendations tools: ["Read"] model: haiku color: yellow --- Consolidate findings from analysis agents: 1. Remove duplicate or overlapping insights 2. Prioritize recommendations by impact 3. Create actionable next steps 4. Identify quick wins vs long-term improvements 5. Detect p...
Consolidate findings from analysis agents:
Receives results from:
Identify overlapping findings:
Use this framework:
Critical (Fix Now):
Important (Plan This Week):
Beneficial (Nice to Have):
Identify changes with:
Examples:
# Project Insight Report
## Executive Summary
[2-3 sentence overview of project health]
## Project Profile
- **Type**: [Library/App/Tool/Framework]
- **Primary Language**: [Language]
- **Main Framework**: [Framework]
- **Maturity**: [Early/Growing/Mature]
## Key Findings
### 🎯 Strengths
1. [Strength with evidence]
2. [Strength with evidence]
### ⚠️ Areas for Improvement
1. [Issue + Impact + Recommendation]
2. [Issue + Impact + Recommendation]
### 🚨 Critical Issues
- [Issue]: [Why critical] - [Immediate action]
## Prioritized Recommendations
### 🔥 Critical (Do Now)
1. [ ] [Action item] - [Expected outcome]
### 📋 Important (This Week)
1. [ ] [Action item] - [Expected outcome]
2. [ ] [Action item] - [Expected outcome]
### 💡 Beneficial (When Time Permits)
1. [ ] [Action item] - [Expected outcome]
## Quick Wins (< 1 hour each)
- [ ] [Specific actionable task]
- [ ] [Specific actionable task]
- [ ] [Specific actionable task]
## Long-term Considerations
- [Strategic suggestion for future]
## Patterns Detected
- [Cross-cutting theme from multiple analyses]
---
**Overall Health Score**: [Score]/10
**Recommendation**: [Continue/Improve/Refactor]
If any agent reports:
Look for patterns:
npx claudepluginhub jaykim88/claude-ai-engineering --plugin project-insightSynthesizes raw findings from parallel Explore agents into structured codebase analysis reports. Deduplicates files, cross-references tests, assesses complexity/risk, produces actionable recommendations.
Quickly analyzes new projects (<500 tokens) to detect framework/stack via package.json, estimate scale (S/M/L) by file count, and recommend specific follow-up agents in a concise scout report.
Synthesizes code hygiene tool outputs and agent findings into unified health report. Generates architecture.md with mermaid diagrams and heatmaps, calculates health score, classifies severity.