Welcome to Clawdbase

The public directory of AI agents and skills with verified identities and trust scores.

AI agents are everywhere. Thousands of tools, assistants, and autonomous systems are published every day - with almost no way to know who built them or whether you can trust them.

Clawdbase fixes that. Every listing carries a trust score based on the builder's verified identity, so you can make informed decisions before you adopt an agent.

How it works

Browse the directory

Visit clawdbase.ai. Every card shows the agent name, creator, trust score, and confidence score at a glance - no account required.

Read the trust scores

Each verified listing shows two numbers: a trust score (how trustworthy the creator is) and a confidence score (how certain we are). Together they tell you how much weight to give the listing.

Explore the lineage

Open any agent page to see the full lineage map - who created it, what it was built on, and whether others have built on top of it.

Use it or list your own

Copy the GitHub link, integrate via the REST API, or connect directly as an MCP tool. If you build agents, register yours to get a verified badge and trust score.

Two types of listings

Verified - The creator registered through MySanctum.ai and put their verified identity behind the agent. These listings show a shield badge, full trust and confidence scores, and an interactive lineage map.

Unverified - Entries discovered from public sources. Searchable and browsable, but without a trust score until the creator registers.

Both appear in the same directory. Verification status is always clearly shown.

Built on inVerus

Trust scores come from inVerus.ai - an open identity verification protocol that analyzes a creator's professional presence across platforms. The score is computed, not self-reported: creators can't set their own score, they can only build the verifiable track record that inVerus uses to calculate one.

Ready to browse? Head to the Getting Started guide, or jump straight to the directory.