https://hermesatlas.com/ Newly added community-built tools, skills, plugins, and integrations for Nous Research's Hermes Agent. Updated daily. en Fri, 01 May 2026 18:50:19 GMT hermesatlas.com/scripts/build-pages.js https://hermesatlas.com/projects/adnw-vinc/hermes-nextcloud https://hermesatlas.com/projects/adnw-vinc/hermes-nextcloud Fri, 01 May 2026 18:48:31 GMT Skills & Skill Registries This project is a Hermes Agent skill that enables direct interaction with self-hosted Nextcloud instances for personal data management. It interfaces with Nextcloud using WebDAV, CalDAV, CardDAV, and the Notes API to perform operations without requiring a browser or third-party sync services. Users can manage files, edit notes, schedule calendar events, track tasks, and search contacts through a command-line interface. The skill prioritizes security by utilizing Nextcloud App Passwords and stori https://hermesatlas.com/projects/plastic-labs/honcho https://hermesatlas.com/projects/plastic-labs/honcho Fri, 01 May 2026 18:48:31 GMT Memory & Context Honcho is an open-source memory library and managed service designed to build stateful agents that maintain long-term context about users, groups, and ideas. It functions as a continual learning system that captures interaction history and applies reasoning to derive insights about entities as they evolve over time. The system provides tools for session management, natural language querying of user state, and automatic conversion of context into model-ready formats like OpenAI's message schema. https://hermesatlas.com/projects/jpalmae/hermeshq https://hermesatlas.com/projects/jpalmae/hermeshq Fri, 01 May 2026 18:48:31 GMT Deployment & Infra HermesHQ is a Docker-first control plane designed to orchestrate and manage multiple Hermes Agent instances from a single web application. It acts as an operational layer over the Hermes execution engine, providing a centralized interface for task dispatching, scheduling, and multi-agent delegation. The platform includes a secrets vault, RBAC, and a runtime ledger to track activity across separate agent workspaces. It also supports managed integrations for AWS Bedrock, OpenAI-compatible endpoint https://hermesatlas.com/projects/longyunfeigu/learn-hermes-agent https://hermesatlas.com/projects/longyunfeigu/learn-hermes-agent Fri, 01 May 2026 18:48:31 GMT Guides & Docs Learn Hermes Agent is a 27-chapter hands-on tutorial for building production-grade autonomous AI agents from scratch using Python. The curriculum guides developers through implementing a core agent loop, tool registration, session persistence with SQLite, and context compression. It covers advanced infrastructure including multi-platform gateways for Telegram and Discord, Model Context Protocol (MCP) integration, and RL-based self-evolution. The project provides runnable reference implementation https://hermesatlas.com/projects/jwangkun/hermes-agent-guide https://hermesatlas.com/projects/jwangkun/hermes-agent-guide Fri, 01 May 2026 18:48:31 GMT Guides & Docs The Hermes Agent White Paper is a comprehensive Chinese-language technical guide designed to bridge the documentation gap for the Hermes Agent framework. It provides a structured 16-book curriculum covering the system's five-layer architecture, three-tier memory system, and the integration of over 47 built-in tools. The guide details practical implementation workflows including local and Docker deployment, MCP protocol automation, and multi-platform connectivity for services like WeChat and Disc https://hermesatlas.com/projects/335234131/agent-browser-mcp https://hermesatlas.com/projects/335234131/agent-browser-mcp Fri, 01 May 2026 18:48:31 GMT Plugins & Extensions Agent-browser-mcp is a Model Context Protocol server that enables AI agents to interact directly with a user's active Chrome browser session rather than a sandboxed environment. It works through a local bridge and a Chrome extension to maintain existing login states, cookies, and open tabs while providing low-level control. The project allows agents to perform page scanning, execute JavaScript, and issue Chrome DevTools Protocol (CDP) commands. It also supports physical hardware interaction, inc https://hermesatlas.com/projects/EfficientContext/ContextPilot https://hermesatlas.com/projects/EfficientContext/ContextPilot Fri, 01 May 2026 18:48:31 GMT Memory & Context ContextPilot is an inference optimization layer designed to accelerate long-context LLM workloads by maximizing KV cache reuse. It maintains a context index to reorder and deduplicate overlapping context blocks, aligning shared content into common prefixes to trigger cache hits. This approach reduces redundant computation and token usage in RAG, agentic workflows, and multi-turn conversations. The tool integrates as a drop-in solution for vLLM, SGLang, llama.cpp, and major cloud APIs without com https://hermesatlas.com/projects/Cranot/super-hermes https://hermesatlas.com/projects/Cranot/super-hermes Fri, 01 May 2026 18:48:31 GMT Skills & Skill Registries Super Hermes is a skill registry for the Hermes Agent that enables models to generate their own analytical frameworks, or prisms, for deep code and document inspection. It works by instructing the agent to write specific thinking instructions before execution, allowing it to derive structural trade-offs and conservation laws rather than just identifying surface-level bugs. The system includes specialized commands like /prism-reflect, which provide a constraint report detailing what the analysis https://hermesatlas.com/projects/iamagenius00/hermes-a2a https://hermesatlas.com/projects/iamagenius00/hermes-a2a Fri, 01 May 2026 18:48:31 GMT Plugins & Extensions hermes-a2a is a plugin for Hermes Agent that implements Google's A2A protocol to enable direct peer-to-peer communication between autonomous agents. It works by injecting external messages into an agent's active session via HMAC-signed webhooks, allowing for instant responses without polling or session cloning. This architecture ensures agents maintain a single shared memory with their user while collaborating on tasks like debugging, status reporting, or peer consultation. The system includes b https://hermesatlas.com/projects/aivrar/portable-hermes-agent https://hermesatlas.com/projects/aivrar/portable-hermes-agent Fri, 01 May 2026 18:48:31 GMT Workspaces & GUIs Portable Hermes Agent is a Windows-native desktop interface for the Hermes Agent framework designed for users who require a zero-install, local-first environment. It operates through a Tkinter-based GUI or CLI, utilizing an embedded Python environment to run without administrative rights or Docker dependencies. The system integrates over 100 tools for tasks such as local model management via LM Studio, media generation through portable TTS and Music servers, and automated image creation with Com https://hermesatlas.com/projects/AMAP-ML/SkillClaw https://hermesatlas.com/projects/AMAP-ML/SkillClaw Fri, 01 May 2026 18:48:31 GMT Skills & Skill Registries SkillClaw is an agentic skill evolution framework that automates the refinement and management of AI agent capabilities based on real-world interactions. It operates as a background process that deduplicates, merges, and improves skill libraries by analyzing session data across multiple agents and devices. The system enables a unified skill repository where experience gained by one agent instance—such as a frontend task—can be cross-pollinated to benefit others. It natively integrates with Herme https://hermesatlas.com/projects/AxDSan/mnemosyne https://hermesatlas.com/projects/AxDSan/mnemosyne Fri, 01 May 2026 18:48:31 GMT Memory & Context Mnemosyne is a local-first memory system designed to provide persistent context for Hermes Agents without relying on cloud services or external databases. It utilizes a SQLite backend with native vector search and full-text search to manage conversations, preferences, and knowledge entirely on-device. The system implements a three-tier BEAM architecture that categorizes data into working, episodic, and scratchpad memory while supporting automatic consolidation through sleep cycles. This approach https://hermesatlas.com/projects/keepnotes-ai/keep https://hermesatlas.com/projects/keepnotes-ai/keep Fri, 01 May 2026 18:48:31 GMT Memory & Context Keep is a reflective memory system for AI agents that enables long-term learning and contextual awareness. It processes diverse inputs—including URLs, codebases, and media files—by summarizing, embedding, and organizing them into a navigable graph of "edge tags." The system automatically surfaces relevant past learnings and commitments during agent sessions to improve decision-making. It supports fully local operation via Ollama or MLX, as well as cloud-hosted and API-based configurations. https://hermesatlas.com/projects/EKKOLearnAI/hermes-web-ui https://hermesatlas.com/projects/EKKOLearnAI/hermes-web-ui Fri, 01 May 2026 18:48:31 GMT Workspaces & GUIs Hermes Web UI is a comprehensive management dashboard for the Hermes Agent ecosystem that provides a graphical interface for AI interaction and system administration. It centralizes multi-platform channel configuration, session management, and usage analytics into a single responsive web environment. Users can monitor token costs, schedule cron jobs, and manage files across local or remote backends like Docker and SSH. The interface also includes specialized tools for group chat orchestration, m https://hermesatlas.com/projects/briancaffey/hermes-otel https://hermesatlas.com/projects/briancaffey/hermes-otel Fri, 01 May 2026 18:48:31 GMT Plugins & Extensions hermes-otel is an OpenTelemetry plugin for the Hermes Agent that enables comprehensive observability for LLM workflows. It automatically captures and exports model invocations, tool calls, and API requests as OTel spans to any OTLP-compatible backend. The plugin supports a wide range of observability platforms including Phoenix, Langfuse, and SigNoz, allowing developers to track traces, metrics, and logs. It features a multi-backend configuration system and a tiered testing suite to ensure relia https://hermesatlas.com/projects/clawvader-tech/hermes-telegram-miniapp https://hermesatlas.com/projects/clawvader-tech/hermes-telegram-miniapp Fri, 01 May 2026 18:48:31 GMT Workspaces & GUIs Hermes Telegram Mini App is a React-based terminal interface that allows users to manage and interact with Hermes agents directly within the Telegram ecosystem. It establishes a secure connection to a local Hermes installation via a web server, providing a mobile-first dashboard for chat, system monitoring, and task scheduling. The application features a terminal-style UI with support for streaming responses, file attachments, and background agent spawning. Security is handled through dual HMAC- https://hermesatlas.com/projects/Codename-11/hermes-relay https://hermesatlas.com/projects/Codename-11/hermes-relay Fri, 01 May 2026 18:48:31 GMT Integrations & Bridges Hermes-Relay is a companion suite that enables remote interaction with a server-deployed Hermes agent through a native Android app and a desktop CLI. It uses a WebSocket Secure (WSS) relay to bridge local devices with remote agents, allowing for chat, terminal access, and cross-device tool execution. The Android application provides mobile device control including screen reading and gestures, while the desktop client allows a remote agent to execute local tools like file searches and screenshots https://hermesatlas.com/projects/stephenschoettler/hermes-lcm https://hermesatlas.com/projects/stephenschoettler/hermes-lcm Fri, 01 May 2026 18:48:31 GMT Memory & Context Hermes-LCM is a context management plugin for the Hermes Agent that prevents information loss during conversation compaction. It replaces standard lossy summarization with a hierarchical Directed Acyclic Graph (DAG) that persists every message in a local SQLite store. This architecture allows the agent to use specialized tools like grep and expand to retrieve specific details from previously compacted history. The system ensures bounded context windows for the model while maintaining an unbounde https://hermesatlas.com/projects/AaronWong1999/hermesclaw https://hermesatlas.com/projects/AaronWong1999/hermesclaw Fri, 01 May 2026 18:48:31 GMT Integrations & Bridges HermesClaw is an integration bridge that allows Hermes Agent, OpenClaw, and OpenCode to operate simultaneously on a single WeChat account. It functions as a centralized iLink poller and proxy server, routing messages to each agent's gateway to prevent the 403 errors and connection drops caused by multiple simultaneous logins. Users can switch between agents or activate multiple brains at once using commands like /hermes, /openclaw, or /three. The tool also enables voice-based coding via an OpenC https://hermesatlas.com/projects/kevinmarmstrong/hermes-claude-code-rc https://hermesatlas.com/projects/kevinmarmstrong/hermes-claude-code-rc Fri, 01 May 2026 18:48:31 GMT Integrations & Bridges This Hermes Agent plugin enables remote management of Claude Code sessions through a Telegram interface. It bridges the gap between mobile devices and local filesystems by allowing users to initiate authenticated remote control sessions on their host machine via slash commands. The plugin captures the resulting session URL and returns it to the user, enabling full access to local tools and databases from the Claude mobile app. Beyond session management, it supports headless command execution and https://hermesatlas.com/projects/pyrate-llama/hermes-ui https://hermesatlas.com/projects/pyrate-llama/hermes-ui Fri, 01 May 2026 18:48:31 GMT Workspaces & GUIs Hermes UI is a glassmorphic web interface that serves as a centralized command center for the self-hosted Hermes Agent. It functions as a single-file React 18 application served via a lightweight Python proxy to provide real-time interaction with the agent's core services. The interface enables users to manage chat sessions, monitor live logs, browse local files, and configure MCP tools or cron jobs through a unified dashboard. It also features a dedicated artifact panel for rendering HTML, SVG, https://hermesatlas.com/projects/garrytan/gbrain https://hermesatlas.com/projects/garrytan/gbrain Fri, 01 May 2026 18:48:31 GMT Memory & Context GBrain is a persistent memory and knowledge graph system designed to prevent AI agents from losing context over time. It utilizes a self-wiring graph architecture that automatically extracts entity relationships and creates typed links without requiring LLM calls. The system combines hybrid search with backlink-boosted ranking to outperform standard vector-only RAG approaches. It integrates with Hermes Agent and OpenClaw via MCP, providing 29 specialized skills for tasks like meeting ingestion, https://hermesatlas.com/projects/nesquena/hermes-webui https://hermesatlas.com/projects/nesquena/hermes-webui Fri, 01 May 2026 18:48:31 GMT Workspaces & GUIs Hermes WebUI is a lightweight web interface for the Hermes Agent, providing a graphical alternative to the standard command-line environment. Built with Python and vanilla JavaScript, it offers a three-panel layout for managing chat sessions, workspace files, and agent settings without requiring a complex build step. The interface maintains full parity with the CLI, allowing users to control models, monitor token usage via a context ring, and manage persistent memory. It facilitates secure remot https://hermesatlas.com/projects/clawshell/clawshell https://hermesatlas.com/projects/clawshell/clawshell Fri, 01 May 2026 18:48:31 GMT Plugins & Extensions ClawShell is a security-privileged process that acts as a safety harness for the Hermes Agent ecosystem by protecting sensitive credentials and PII. It functions as a proxy between the agent and upstream LLM providers, swapping virtual keys for real ones stored in a restricted directory to ensure the agent never has direct access to credentials. The system performs Data Loss Prevention (DLP) scanning to redact or block sensitive information in request and response bodies. Additionally, it provid https://hermesatlas.com/projects/dodo-reach/hermes-desktop https://hermesatlas.com/projects/dodo-reach/hermes-desktop Fri, 01 May 2026 18:48:31 GMT Workspaces & GUIs Hermes Desktop is a native macOS application that provides a localized workspace for managing Hermes Agent hosts over SSH. It connects directly to remote environments to manage sessions, skills, and cron jobs without requiring a gateway API, local file mirroring, or remote daemon. The app features a multi-tabbed terminal and profile-aware views that surface usage trends and canonical file states directly from the host. This allows users to maintain a persistent, native interface for multi-agent https://hermesatlas.com/projects/ksimback/hermes-ecosystem https://hermesatlas.com/projects/ksimback/hermes-ecosystem Fri, 01 May 2026 18:48:31 GMT Guides & Docs Hermes Atlas is a community-curated directory and ecosystem map for the Hermes Agent by Nous Research. It aggregates over 80 quality-filtered repositories, including skills, plugins, and deployment templates, using a vanilla JavaScript frontend and Vercel serverless functions. The platform features a RAG-powered chatbot that answers technical questions grounded in 27 research files using hybrid retrieval and OpenRouter LLMs. Users can search and filter the ecosystem while viewing live GitHub sta https://hermesatlas.com/projects/xaspx/hermes-control-interface https://hermesatlas.com/projects/xaspx/hermes-control-interface Fri, 01 May 2026 18:48:31 GMT Workspaces & GUIs Hermes Control Interface is a self-hosted web dashboard designed to simplify the management of the Hermes AI agent stack. It utilizes a Node.js and Express backend with a Vanilla JS and Vite frontend to provide a browser-based administrative layer. The tool enables users to monitor system health, manage multiple agent profiles, and analyze token usage and costs. It also includes integrated utilities for file editing, terminal access, and cron job scheduling. https://hermesatlas.com/projects/beardthelion/hermes-skill-distillation https://hermesatlas.com/projects/beardthelion/hermes-skill-distillation Fri, 01 May 2026 18:48:31 GMT Forks & Derivatives This project provides a skill distillation pipeline designed to fine-tune Hermes 4 models using real-world agentic trajectories. It captures tool-use sessions from the Hermes agent and converts them into high-quality training data via a specialized environment called RealWorldTaskEnv. The system automatically scores these trajectories based on task completion, efficiency, and error recovery to create SFT-ready datasets or live RL training loops. By integrating with the Atropos framework, it allo https://hermesatlas.com/projects/kaminocorp/hermes-alpha https://hermesatlas.com/projects/kaminocorp/hermes-alpha Fri, 01 May 2026 18:48:31 GMT Forks & Derivatives Hermes Alpha is an experimental framework designed to test if a stock AI agent can autonomously bootstrap and manage a bug bounty system from first principles. The architecture utilizes a two-agent loop where a persistent 'Overseer' agent builds, deploys, and refines an ephemeral 'Hunter' agent tasked with finding software vulnerabilities. By providing the agent with only a Linux terminal and an identity document rather than custom tools, the project evaluates the self-improvement capabilities o https://hermesatlas.com/projects/nativ3ai/hermes-agent-camel https://hermesatlas.com/projects/nativ3ai/hermes-agent-camel Fri, 01 May 2026 18:48:31 GMT Forks & Derivatives This project is a specialized fork of the Hermes Agent that integrates CaMeL trust-boundary principles to defend against prompt-injection attacks during autonomous execution. It implements a security layer adapted from Google Research that monitors or enforces policy decisions when the agent interacts with untrusted contexts. Users can explicitly opt into these protections using specific CLI flags to block unauthorized sensitive actions without altering the core Hermes runtime. The agent retains