# agentsite-static-version=12473b6e-ed18-472a-8b5a-1e2056ac7a6a # agentwikis.com > AI-optimized mirror of agentwikis.com containing 50 pages totalling 47,121 words of clean markdown content, structured data, and semantic HTML. Original source: https://agentwikis.com. Last updated: 2026-09-16T04:05:25.169Z. Each page is available as HTML (with JSON-LD structured data) and Markdown (text-only, ideal for LLMs and RAG). ## Homepage - [Agent Wikis — Maintained knowledge bases for humans and AI agents](/content/site-root.html): Agent Wikis compiles messy, fast-moving sources into maintained, structured wikis that humans browse and AI agents read natively as Markdown and llms.txt. (1,458 words) ## Articles & Blog Posts - [raw/hermes/wiki/log-md.html](/content/raw/hermes/wiki/log-md.html) (5,259 words) - [CLI Reference — hermes Terminal Commands — Hermes — Agent Wikis](/content/wiki/hermes/wiki/concepts/cli-reference-md.html): The full surface of hermes commands you run from a shell. In-chat slash commands (/model, /new, /background … typed inside a CLI session or a messaging cha… (3,910 words) - [wiki/hermes/index-json.html](/content/wiki/hermes/index-json.html) (1,396 words) - [Model Switching & Credential Pools — Hermes — Agent Wikis](/content/wiki/hermes/wiki/concepts/model-switching-md.html): Hermes is deliberately provider-agnostic. It speaks to every major hosted inference API and every major local runtime through a uniform provider abstractio… (1,776 words) - [raw/hermes/wiki/index-md.html](/content/raw/hermes/wiki/index-md.html) (1,077 words) - [raw/buzz/wiki/concepts/git-on-nostr-md.html](/content/raw/buzz/wiki/concepts/git-on-nostr-md.html) (948 words) - [wiki/hermes/llms-txt.html](/content/wiki/hermes/llms-txt.html) (1,023 words) - [Approval System — Hermes — Agent Wikis](/content/wiki/hermes/wiki/concepts/approval-system-md.html): The approval system is Hermes' guardrail against catastrophic shell commands. Every terminal and process invocation is matched against a hand-curated list… (1,206 words) - [Onchain Workflows — Hermes — Agent Wikis](/content/wiki/hermes/wiki/concepts/onchain-workflows-md.html): Onchain workflows in Hermes Agent are any task that reads from or writes to a (1,495 words) - [Memory System — Hermes — Agent Wikis](/content/wiki/hermes/wiki/concepts/memory-system-md.html): Hermes Agent's memory subsystem is a two-layer architecture that separates the always-on built-in store from a pluggable provider interface. The built-in l… (1,481 words) - [Migration from OpenClaw — Hermes — Agent Wikis](/content/wiki/hermes/wiki/concepts/migration-from-openclaw-md.html): hermes claw migrate is the official, in-tree command for moving an existing OpenClaw setup over to Hermes Agent. It transfers the user-facing parts of an O… (1,336 words) - [Auxiliary Models — Dedicated Providers for Built-in Side Tasks — Hermes — Agent Wikis](/content/wiki/hermes/wiki/concepts/auxiliary-models-md.html): Hermes routes a set of internal "side tasks" through a separate, user-configurable model path — the auxiliary model system. Instead of burning your main ag… (1,620 words) - [Cron & Scheduling — Hermes — Agent Wikis](/content/wiki/hermes/wiki/concepts/cron-scheduling-md.html): Hermes' cron system lets the agent run itself on a schedule — a persistent, in-process scheduler that turns any natural-language prompt into a recurring or… (1,476 words) - [Skills System — Hermes — Agent Wikis](/content/wiki/hermes/wiki/concepts/skills-system-md.html): Skills are Hermes Agent's procedural memory — markdown documents that the agent loads on demand to learn how to do specific tasks. Each skill is a self-con… (1,110 words) - [MCP Integration — Hermes — Agent Wikis](/content/wiki/hermes/wiki/concepts/mcp-integration-md.html): Model Context Protocol (MCP) integration in Hermes has three distinct directions: (1) Hermes consumes external MCP servers as additional tools, (2) Hermes… (1,313 words) - [Self-Improvement Loop (GEPA) — Hermes — Agent Wikis](/content/wiki/hermes/wiki/concepts/self-improvement-loop-md.html): The self-improvement loop is Hermes' headline differentiator: a closed cycle in which the agent observes its own behavior, identifies what failed, packages… (1,623 words) - [Local Models & Airplane Mode — Hermes — Agent Wikis](/content/wiki/hermes/wiki/concepts/local-models-airplane-mode-md.html): Hermes can run 100% locally — no API keys, no cloud round-trips, no telemetry leaving the box — but there is no single --airplane-mode flag that flips this… (1,219 words) - [Messaging Gateway — Hermes — Agent Wikis](/content/wiki/hermes/wiki/concepts/messaging-gateway-md.html): The messaging gateway is a single long-running background process (gateway/run.py) that multiplexes platform adapters into one shared agent loop. As of v0.… (920 words) - [Web Search & Extract — Hermes — Agent Wikis](/content/wiki/hermes/wiki/concepts/web-search-extract-md.html): Hermes exposes two model-callable web tools: websearch returns ranked results, while webextract fetches readable content from URLs. Search and extraction c… (637 words) - [Persistent Goals (/goal) & Verification — Hermes — Agent Wikis](/content/wiki/hermes/wiki/concepts/goals-and-verification-md.html): /goal gives Hermes a standing objective that survives across turns: after every turn a lightweight judge model checks whether the goal is satisfied, and if… (1,040 words) - [Mixture of Agents (MoA) — Hermes — Agent Wikis](/content/wiki/hermes/wiki/concepts/mixture-of-agents-md.html): Mixture of Agents is a virtual model provider (added v0.18.0, graduating from an earlier toggle mode). Each named MoA preset appears as a selectable model… (1,037 words) - [ML Research Pipeline — Hermes — Agent Wikis](/content/wiki/hermes/wiki/concepts/ml-research-pipeline-md.html): Hermes ships a complete ML and reinforcement-learning research toolchain inside the agent — not as an optional add-on, but as a first-class set of tools, s… (836 words) - [Deployment Backends — Hermes — Agent Wikis](/content/wiki/hermes/wiki/concepts/deployment-backends-md.html): A deployment backend (also called the terminal backend in Hermes parlance) is the execution environment that the agent's terminal and process tools run com… (903 words) - [Hermes Agent Knowledge Base — Hermes — Agent Wikis](/content/wiki/hermes/readme-md.html): LLM-maintained research KB on Hermes Agent — used as the research backbone for YouTube videos and the "Complete Hermes Agent Masterclass" (149 words) - [SSH / VPS Backend — Hermes — Agent Wikis](/content/wiki/hermes/wiki/entities/backend-ssh-md.html): The SSH backend runs every Hermes shell command on a remote server over SSH — a $5 VPS, a remote GPU box, or any machine you can reach with a key. It is th… (547 words) - [Bot Mode & Teammate Rooms — Hermes — Agent Wikis](/content/wiki/hermes/wiki/concepts/bot-mode-md.html): Bot Mode is Hermes' persistent multi-agent collaboration surface. Introduced as the bundled hermes-bots plugin in v0.20.3, it is built into the desktop app… (614 words) - [Automation Blueprints — Hermes — Agent Wikis](/content/wiki/hermes/wiki/entities/automation-blueprints-md.html): Automation Blueprints (added v0.17.0, #41309) let you schedule an automation by picking a named template and answering the questions it asks, instead of wr… (547 words) - [Subagents & Delegation — Hermes — Agent Wikis](/content/wiki/hermes/wiki/concepts/subagents-delegation-md.html): Delegation in Hermes is the ability for a running agent to spawn child AIAgent instances via the delegatetask tool, each with an isolated context window, i… (563 words) - [Docker Container Backend — Hermes — Agent Wikis](/content/wiki/hermes/wiki/entities/backend-docker-md.html): The Docker terminal backend runs every Hermes shell command and skill script inside a disposable container instead of on the host. It gives you process- an… (420 words) - [Daytona Managed Workspace Backend — Hermes — Agent Wikis](/content/wiki/hermes/wiki/entities/backend-daytona-md.html): The Daytona backend runs Hermes shell commands in a Daytona managed serverless workspace — a cloud dev container that hibernates when idle and resumes on t… (434 words) - [Configuration Reference — ~/.hermes/ Layout & Management — Hermes — Agent Wikis](/content/wiki/hermes/wiki/concepts/configuration-reference-md.html): After updating Hermes, run hermes config check — new versions add options; migrate fills them interactively. Easiest path to a working config.yaml on a fre… (367 words) - [Singularity / Apptainer HPC Backend — Hermes — Agent Wikis](/content/wiki/hermes/wiki/entities/backend-singularity-md.html): The Singularity / Apptainer backend runs Hermes shell commands inside a Singularity (now Apptainer) container — the container runtime of choice on HPC clus… (397 words) - [raw/mcp/wiki/concepts/transports-md.html](/content/raw/mcp/wiki/concepts/transports-md.html) (842 words) - [LLM Wiki — gbrain — Agent Wikis](/content/wiki/gbrain/readme-md.html): An open-source template for building LLM-powered knowledge bases, following Andrej Karpathy's "LLM Wiki" pattern. (803 words) - [LLM Wiki — Pi Agent — Agent Wikis](/content/wiki/pi-agent/readme-md.html): An open-source template for building LLM-powered knowledge bases, following Andrej Karpathy's "LLM Wiki" pattern. (805 words) - [Local Terminal Backend — Hermes — Agent Wikis](/content/wiki/hermes/wiki/entities/backend-local-md.html): The local terminal backend runs Hermes Agent commands directly on the host machine. It is the default backend — used unless you explicitly configure anothe… (241 words) - [autonovel — Hermes — Agent Wikis](/content/wiki/hermes/wiki/entities/autonovel-md.html): autonovel (github.com/NousResearch/autonovel, 467 stars per get-hermes.ai, 2026-04-12) is an official entities/nous-research project: an autonomous novel-w… (232 words) - [Modal Serverless Backend — Hermes — Agent Wikis](/content/wiki/hermes/wiki/entities/backend-modal-md.html): The Modal backend runs Hermes shell commands and skills on Modal's serverless container platform — ephemeral compute that hibernates when idle and resumes… (253 words) - [Skills — Agent Wikis](/content/skills/index.html): Agent skills developed over months of real Hermes Agent work — not evals — refined for release. Thin SKILL.md disciplines that pull their facts from maintained Agent Wikis over MCP. Included with Pro. (719 words) - [For agents — Agent Wikis](/content/for-agents/index.html): How AI agents consume Agent Wikis: llms.txt, raw Markdown, content negotiation, the MCP server, and trust semantics (scope, confidence, provenance). (633 words) - [raw/mcp/wiki/concepts/lifecycle-md.html](/content/raw/mcp/wiki/concepts/lifecycle-md.html) (902 words) - [Why Agent Wikis? — Agent Wikis](/content/why-wikis/index.html): Measured evidence: compiled, maintained wikis give agents 89% correct answers vs 48% with web search, with 7% vs 48% hallucination, ~3× cheaper and faster. (487 words) - [wiki/pi-agent/index-json.html](/content/wiki/pi-agent/index-json.html) (257 words) - [Pro — Agent Wikis](/content/pro/index.html): Agent Wikis Pro: XL wiki editions with full reference depth for AI agents. $9.99/month. (245 words) - [Beginner's Guide to Agent Wikis — Videos — Agent Wikis](/content/videos/beginners-guide/index.html): What an LLM wiki is, why agents answer better from one, and how to connect Agent Wikis to your own agent in a single command. Start here. (48 words) - [Videos — Agent Wikis](/content/videos/index.html): Short walkthroughs of Agent Wikis in practice: connecting the MCP server, working with XL wikis, installing skills, and running a workflow end to end. (112 words) - [wiki/pi-agent/llms-txt.html](/content/wiki/pi-agent/llms-txt.html) (207 words) - [raw/hermes/readme-md.html](/content/raw/hermes/readme-md.html) (97 words) ## About Pages - [About — Agent Wikis](/content/about/index.html): About Agent Wikis — maintained knowledge bases for humans and AI agents. (101 words) ## Resources - [Full Page Index](/index.html): Browse all cached pages with rich metadata - [About This Cache](/about.html): Methodology, technical details, and usage guidelines - [XML Sitemap](/sitemap.xml): Machine-readable sitemap for crawler discovery - [Robots.txt](/robots.txt): Crawler directives - [AgentSite Network](https://agentsite.network/network.html): Public index of AgentSites and their machine-readable resources