Repository Radar - PR#18
This issue's picks are up + 174% combined since it ran (115k → 315k). DeepResearch has been the runaway (1322×), while spec-kit piled on the most stars outright (+100k). GrowthBook has barely budged (+18%). In the past month, spec-kit (+8k) and chrome-devtools-mcp (+3k) are still climbing. In the understory, GrowthBook is shipping harder than its stars suggest (understory score +33).
- DeepResearch132k(15 → 20k)
- chrome-devtools-mcp665(7k → 49k)
- spec-kit447(29k → 129k)
- RAG-Anything310(7k → 23k)
- HumanLayer / CodeLayer235(5k → 11k)
- LLaMA-Factory125(60k → 74k)
- GrowthBook118(7k → 8k)
star trajectories since covered are charted per repo below · compare them all →
// above the radar
- 📚GrowthBook ACTIVE
Open Source Feature Flags and A/B Testing
8k now 7k when covered + 1k (+18%)
// on the radar
- 📋spec-kit ACTIVE
Spec-Driven Development Toolkit
129k now 29k when covered + 100k (+347%) - 🦙LLaMA-Factory ACTIVE
Unified Fine-Tuning of 100+ LLMs & VLMs
74k now 60k when covered + 15k (+25%) - 🕵️DeepResearch ACTIVE
Long-Horizon Research Agent
20k now 15 when covered + 20k (+132100%)
// below the radar
- 🔧chrome-devtools-mcp ACTIVE
Expose Chrome DevTools safely to coding agents
49k now 7k when covered + 42k (+565%) - 📊RAG-Anything ACTIVE
All-in-One Multimodal RAG System
23k now 7k when covered + 16k (+210%) - 🧠HumanLayer / CodeLayer ACTIVE
Open IDE for orchestrating Claude Code across repos
11k now 5k when covered + 6k (+135%)
This is the archived scan with live tracking. Read the full sweep on Substack ↗