# GARL Protocol

> Use this tool when you need to establish trust and reputation among AI agents interacting with each other, solving problems of unreliable task delegation and insecure data exchange. The GARL Protocol provides a robust interface for registering agents, logging actions, and querying trust scores, with inputs including agent registrations and action logs, and outputs of 5-dimensional trust scores and reliability queries. It is ideal for use cases requiring secure and trustworthy AI agent collaborations, with support for integration via REST API, Python and JavaScript SDKs, and Google A2A protocol.

Canonical page: https://skillsregistry.net/skills/garl-protocol  
JSON: https://api.skillsregistry.net/v1/skills/garl-protocol

## Description

Provides a trust and reputation protocol for AI agent interactions. Agents can register, log actions, build 5-dimensional trust scores using exponential moving averages, and query each other's reliability before delegating tasks. Includes cryptographic signing with ECDSA, immutable audit trails, smart routing by category and certification tier, and ERC-8004 compatibility. Supports integration via REST API, Python and JavaScript SDKs, and Google A2A protocol.

## Trust

- **Trust score (0–1):** 0.00
- **Verification tier:** scanned
- **Last scanned:** 2026-09-19

## Facts

- **Version:** 1.0.0
- **Skill type:** atomic
- **Execution layer:** mcp-remote
- **Runtime environment:** api
- **Category:** finance
- **Updated:** 2026-09-19

## Source

- **Source listing:** [PulseMCP](https://www.pulsemcp.com/servers/garl-protocol)
- **Repository:** <https://github.com/garl-protocol/garl>

## Use it

Resolve this record through the SkillsRegistry MCP server (no auth, read-only):

```
claude mcp add --transport http --scope user skillsregistry https://api.skillsregistry.net/mcp
```

```json
{
  "jsonrpc": "2.0",
  "id": 1,
  "method": "tools/call",
  "params": {
    "name": "get_skill",
    "arguments": {
      "slug": "garl-protocol"
    }
  }
}
```

REST: `GET https://api.skillsregistry.net/v1/skills/garl-protocol` · pull for local use: `GET https://api.skillsregistry.net/v1/skills/garl-protocol/pull`

---
SkillsRegistry indexes agent skills from public registries and GitHub. Skills we have analysed are scanned with Circle-IR and scored on six dimensions; each listing states its scan coverage. More: https://skillsregistry.net/llms.txt