42-evey
evey-validate
Hallucination detection for Hermes agents using regex patterns and LLM scoring (0-10).
Repository last updated Jul 18, 2026
Overview
Evey Validate is a plugin for the Hermes Agent ecosystem that detects hallucinations in agent outputs. It combines regex-based pattern matching with LLM-based scoring on a 0–10 scale to identify potentially incorrect or fabricated information.
The plugin is designed for developers building autonomous AI agents who need a lightweight, configurable way to validate responses before passing them downstream. By catching hallucinations early, you can improve the reliability and trustworthiness of your agent workflows.
Evey Validate integrates directly into the Hermes Agent plugin system, making it easy to add hallucination detection without modifying core agent logic. It is open-source under the MIT license and written in Python.
Highlights
- Detects hallucinations in agent outputs
- Uses regex patterns for fast initial filtering
- LLM-based scoring from 0 (low) to 10 (high)
- Lightweight Python plugin for Hermes Agent
- Configurable to suit different use cases
Features
Regex Pattern Matching
Apply custom regex patterns to flag suspicious phrases or contradictions in agent responses.
LLM Scoring
Score each output on a 0–10 scale using an LLM to assess hallucination likelihood.
Hermes Agent Integration
Plug directly into the Hermes Agent plugin system for seamless addition to agent pipelines.
Installation
Commands are reproduced exactly as published by the project maintainers. Always check the upstream repository for the current instructions.
git clone https://github.com/42-evey/hermes-plugins.git
cp -r hermes-plugins/evey-* ~/.hermes/plugins/
cp hermes-plugins/evey_utils.py ~/.hermes/plugins/
# Restart hermes-agentRequirements
- Hermes Agent installed
- Python 3.x
How it's built
Evey Validate is written in Python and designed as a plugin for the Hermes Agent ecosystem. It uses a two-stage detection process: first, regex patterns are applied to quickly identify common hallucination indicators; second, an LLM evaluates the output and assigns a score from 0 (no hallucination) to 10 (highly likely hallucination).
The plugin is configured via standard Hermes Agent plugin mechanisms. No specific repository layout or version information is provided, but the code is available on GitHub under the MIT license.
Source
This summary was written for Neura Market from the project's own repository. Install commands, identifiers and licence details are reproduced unchanged. The 42-evey/evey-validate repository is the authoritative source and the place to file issues or contribute.
Category: Observability & Telemetry · Licensed under MIT