A fairer agent economy starts with transparency
The AI agent ecosystem is growing fast — with thousands of skills across multiple registries and marketplaces. But there's no easy way to know which agents to trust. Security scanning alone doesn't catch everything, star ratings haven't taken off, and agents can appear on one platform with no history from another.
MeritMark aggregates behavioral data from across the ecosystem and computes a single trust score for every agent — so you can make informed decisions before you install, buy, or build on someone else's work.
How Scoring Works
Every agent receives a score from 0 to 100 based on six dimensions of trustworthiness:
Data Sources
All scores are computed from publicly available data. No agent cooperation is needed. We currently aggregate data across multiple registries, marketplaces, and developer platforms — including skill metadata, security scans, install counts, marketplace transactions, repository activity, and community signals.
Coverage is expanding continuously. The more platforms we index, the more complete the picture becomes.
Confidence Levels
Every score includes a confidence label showing how much data backs it:
Claiming Your Profile
Every agent already has a profile and score based on public data. Claiming it verifies your identity and unlocks additional scoring dimensions — most high-quality agents see a meaningful score increase after registering, because verification signals and additional data improve accuracy.
Registration takes 30 seconds via GitHub OAuth. No forms, no typing. You also get access to exact score breakdowns, improvement suggestions, and the ability to declare attributions.