The Big Picture
Star counts are a weak guide; look at contributor density, cross-project contributors, and early retention (30–90 days) to judge a toolkit's real community strength.
ON THIS PAGE
The Evidence
Popular repositories can be full of hype: rapid star growth often does not translate into real contributors. Lower-profile projects sometimes have a deeper, more committed developer base. One toolkit (LangChain) acts like a shared backbone, drawing most cross-project contributors, and community activity drops fastest in the first 30 days after someone first contributes before stabilizing around 90 days.
Not sure where to start?Get personalized recommendations
Data Highlights
1AutoGPT gained 111,967 stars in one month but converted to fewer than 9 contributors per 1,000 stars (low contributor density).
2LangChain achieved about 41 contributors per 1,000 stars, showing stronger conversion from visibility to actual contributors.
382.5% of contributors who work across multiple agent projects contributed to LangChain, making it the dominant shared infrastructure.
What This Means
Engineering managers and platform teams choosing a toolkit: use these metrics to avoid picking a trendy but shallow project. Tool maintainers and community builders: focus on converting early interest into committed contributors and cross-project connections. Researchers and evaluators tracking multi-agent trust and agent track records will find more reliable signals in contribution patterns than star counts.
Ready to evaluate your AI agents?
Learn how ReputAgent helps teams build trustworthy AI through systematic evaluation.
Learn MoreYes, But...
Metrics were collected from 15 major open-source agent toolkits between late 2022 and early 2026, so findings may not generalize to private repos or non-GitHub ecosystems. Contributor density and cross-project activity favor projects that integrate widely, which can disadvantage niche but high-quality tools. Retention patterns describe contribution activity and not necessarily code quality or runtime reliability—use these signals alongside technical evaluation and testing.
Methodology & More
Researchers tracked 15 open-source AI agent toolkits from late 2022 to early 2026, mining public GitHub data: 808,042 stars, 73,997 pull requests, 86,241 commits, and profiles from 987,330 users. They defined simple, practical metrics: contributor density (contributors per 1,000 stars), cross-ecosystem contribution (who contributes to multiple projects), and short-term retention (how long contributors stay active after their first contribution). These measures were chosen to separate headline popularity from actual community depth.
Key findings show that raw star counts can be misleading because of hype and inorganic attention. High-visibility projects sometimes convert poorly to contributors, while lower-profile repos can have higher contributor density, signaling deeper adoption. One project functions as a shared infrastructure hub, capturing most cross-project contributors and amplifying its ecosystem role. Contribution activity falls off steeply within the first 30 days and tends to stabilize by about 90 days, so early contributor engagement is a critical window. For practical decisions, use contributor density, cross-project engagement, and early retention as part of your rubric when evaluating which toolkit to adopt or support, and combine these with hands-on testing for reliability and safety.
Avoid common pitfallsLearn what failures to watch for
Credibility Assessment:
All authors lack well-known affiliations; one author h-index=8 (low). Venue is arXiv with no citations — limited signals, so emerging/limited credibility.