Curated, install-ready skills for Hermes Agent (v0.17.0). 72 built-in + 101 optional bundled skills + 85 community skills, plugins & tools for the self-improving AI agent from Nous Research. Cross-compatible with Claude Code, OpenClaw, Cursor, and Windsurf.
A curated, install-ready directory for the Hermes Agent ecosystem — the self-improving AI agent from Nous Research. Covers the 82 built-in skills and 117-skill optional catalog that ship with Hermes v0.20.6, plus 169 community skills, plugins, agent profiles, memory providers, surfaces, and tools vetted for quality.
Hermes is the only agent with a real learning loop. It writes its own skills from your workflows, searches its own past conversations, and runs anywhere — a $5 VPS, a GPU cluster, serverless, or your laptop. But the agent is only as powerful as the skills you give it. This list is the shortcut. Pick three, install in a minute, and your agent is twice as useful by tonight.
⭐ Featured Skill
youtube-full : Get YouTube transcripts, search videos, browse channels, and extract playlists from any AI agent.
The fastest way to feel what Hermes can do. Install it, then ask: *"Summarize the last 5 videos from Lex Fridman."* Your agent fetches the transcripts, summarizes them, and remembers what you watched, all in one prompt. No Google API key, and nothing to install or maintain. Powered by TranscriptAPI. Works in Hermes, Claude, OpenClaw and other agent runtimes.
New to Hermes? Don't try to install everything at once. Here's the three-step path from zero to "wait, my agent can do that?":
1
Get Hermes running — Follow the official quickstart. 10 minutes from clone to first conversation. Hermes ships with 82 built-in skills — plus a 117-skill optional catalog — the day you install it, so even before you add anything from this list, you can already do a lot.
Then ask: *"Get the transcript for this video and summarize it: [paste any YouTube URL]"*. The "I get it" moment lands in 30 seconds. From there try: *"Find the top 5 videos about quantum computing and tell me what they agree on."*
3
Pick a workflow upgrade — Add hermes-workspace for a full GUI, or SkillClaw so your skill library auto-evolves while you work.
Browse the categories below. Every entry is tagged so you know what you're getting:
Tag
What it means
built-in
Ships with Hermes — already installed when you run hermes
production
Stable, documented, actively maintained — safe to build on
A short hand-picked list to get you started. If you install nothing else from this page, install these.
*Selection is editorial; the order is not. Picks are listed most-starred first, so the ranking is something you can check rather than something you have to trust. Counts re-pulled from the GitHub API 2026-08-24 09:23 UTC; on a list this size they drift within hours, so treat the figure as of that timestamp and the ordering as the durable claim.*
Twenty-three battle-tested skills from the Total TypeScript creator. The grill-me and grill-with-docs skills force the agent to interview you before writing code — the single best fix for "the agent didn't do what I wanted." tdd enforces red-green-refactor. improve-codebase-architecture rescues codebases that have become balls of mud. wayfinder maps work too big for one session onto decision tickets. Built for real engineering work, not vibe coding. Pick the ones you want with the installer's interactive picker.
Web-based workspace with chat, terminal, memory browser, skills manager, and inspector. The most complete GUI for Hermes. Built during Nous Hackathon 2026. Pairs well with everything else on this list.
Open-source companion that auto-evolves, deduplicates, and improves your skill library from real session data. Sits on top of Hermes's built-in skill creation and adds a post-task evolution loop. Native Hermes integration via ~/.hermes/skills, with safety flows (skillclaw doctor hermes / skillclaw restore hermes).
If your agent can't read YouTube, half the internet is invisible to it. This skill gives Hermes transcript extraction, channel browsing, search, and playlist parsing — no Google API key, and nothing to install or maintain. Powered by TranscriptAPI. Works in Hermes, Claude, OpenClaw and other agent runtimes.
Deepfake detection for agents that ingest user-submitted media. Detects AI-generated audio, image, video, and text. Traces audio source (ElevenLabs, Resemble, etc.), applies invisible watermarks for provenance, and verifies speaker identity. The first thing to install if your agent reads the public internet.
📦 Built-in Skills (Ships with Hermes)
Hermes ships with 82 built-in skills out of the box, across 14 categories — loaded and ready the moment you run hermes. You don't install these.
>
This is what you already have before you add anything. Most "I need a skill for X" questions are answered here first — and the optional catalog below covers 117 more you can switch on.
🍎 Apple (4)
apple-notes — Manage Apple Notes via memo CLI: create, search, edit.
apple-reminders — Apple Reminders via remindctl: add, list, complete.
findmy — Track Apple devices/AirTags via FindMy.app on macOS.
imessage — Send and receive iMessages/SMS via the imsg CLI on macOS.
🤖 Autonomous Ai Agents (6)
claude-code — Delegate coding to Claude Code CLI (features, PRs).
codex — Delegate coding to OpenAI Codex CLI (features, PRs).
computer-use — Drive the desktop in the background without stealing focus.
hermes-agent — Use, configure, theme, extend, and orchestrate Hermes Agent.
merge-reconciler — Neutral third-party resolution of agent merge conflicts.
opencode — Delegate coding to OpenCode CLI (features, PR review).
🎨 Creative (16)
architecture-diagram — Dark-themed SVG architecture/cloud/infra diagrams as HTML.
simplify-code — Parallel 4-agent cleanup of recent code changes.
spike — Throwaway experiments to validate an idea before build.
systematic-debugging — 4-phase root cause debugging: understand bugs before fixing.
test-driven-development — TDD: enforce RED-GREEN-REFACTOR, tests before code.
🧩 Optional Skills (Bundled with Hermes)
Beyond the always-on built-ins, Hermes bundles a 117-skill optional catalog — shipped in the repo but off by default, so your context stays lean until you need them. Browse [optional-skills/](https://github.com/NousResearch/hermes-agent/tree/main/optional-skills) and enable the ones your workflow calls for.
>
This is where the deep, domain-specific power lives: 31 MLOps skills (training, serving, vector DBs, interpretability), a full finance-modeling suite (DCF, LBO, merger, 3-statement), plus OSINT, blockchain, payments, and more.
🤖 Autonomous Ai Agents (5)
antigravity-cli — Operate the Antigravity CLI (agy): plugins, auth, sandbox.
blackbox — Delegate coding tasks to the Blackbox AI multi-model CLI.
Skills, plugins, and integrations built by the Hermes community. Tagged for maturity. Click through for install instructions per project.
📺 Media & Transcripts
hermes-spotify-skill by Alexeyisme — Spotify control for headless Linux and Raspberry Pi. The only Linux-native Spotify skill in the ecosystem. [beta]
youtube-skills by therohitdas — Transcripts, search, channels, playlists. Cross-agent: Hermes, Claude, OpenClaw and other agent runtimes. Powered by TranscriptAPI. [production]
🔍 Search & Research
consensus-mcp-hermes by ahmdngi — Connect 200M+ peer-reviewed research papers to Hermes via the Consensus MCP server. OAuth setup guide for headless environments with mcp-remote bridge. [beta]
deep-research by moonlight-lupin — Autonomous Think→Search→Extract→Synthesize→Stop loop that produces cited research reports. [production]
hermes-web-search-plus by robbyczgw-cla — Multi-provider web search with intelligent routing across Serper, Tavily, Exa, and more. Replaces built-in search with better quality + source diversity. [beta]
Not Human Search by unitedideas — MCP server for discovering other MCP servers. Indexes 8,600+ agent-friendly sites with agentic scoring. Lets Hermes find new tools on its own. [production]
Scavio by scavio-ai — Structured web data from 31 platforms through one hosted MCP server: search, retail, social, property and SEC filings, plus five research workflow skills. Credit-metered, with 50 free credits on signup. [beta]
📈 Marketing & Growth
hermes-startup by 33hodl — Walks a personal profile into a ranked idea shortlist and a tool plan for a first paying customer. No income claims. [beta]
notfair-plugin by NoWork Studio — 40+ host-agnostic skills for SEO, GEO, Google Ads and Meta Ads, with approval-gated MCP actions on connected accounts. [beta]
💻 Dev & Skill Authoring
🔥 mattpocock/skills — Engineering bundle
Install all at once: npx skills@latest add mattpocock/skills — pick which skills you want at install time.
ask-matt by mattpocock — Router over the rest of the pack: describe your situation, get told which skill or flow fits. [production]
code-review by mattpocock — Reviews changes since a fixed point along two axes at once — does it follow the repo's standards, and does it match what the spec asked for. [production]
codebase-design by mattpocock — Shared vocabulary for designing deep modules: where a seam goes, what makes code testable and AI-navigable. [production]
diagnosing-bugs by mattpocock — Disciplined diagnosis loop for hard bugs and performance regressions: reproduce, minimise, hypothesise, instrument, fix. [production]
domain-modeling by mattpocock — Pins down a project's ubiquitous language and records architectural decisions as you go. [production]
git-guardrails-claude-code by mattpocock — Blocks dangerous git commands — push, reset --hard, clean, branch -D — before they execute. Hooks-based. [production]
grill-with-docs by mattpocock — Relentless interview that sharpens a plan and writes the ADRs and glossary entries as it goes. The most popular skill in the pack. [production]
implement by mattpocock — Builds a piece of work from an existing spec or set of tickets, rather than from a conversation. [production]
improve-codebase-architecture by mattpocock — Scans for deepening opportunities, presents them as a visual HTML report, then grills through whichever you pick. [production]
migrate-to-shoehorn by mattpocock — Migrates test files from as type assertions to @total-typescript/shoehorn. [production]
prototype by mattpocock — Builds a throwaway prototype to answer one design question, then throws it away. [production]
research by mattpocock — Investigates a question against primary sources and commits the findings as Markdown in the repo. [production]
scaffold-exercises by mattpocock — Creates exercise directory structures with sections, problems, solutions and explainers. [production]
setup-matt-pocock-skills by mattpocock — Run once first: configures the repo's issue tracker, triage labels and domain doc layout for the rest of the pack. [production]
setup-pre-commit by mattpocock — Sets up Husky pre-commit hooks with lint-staged, Prettier, type checking and tests. [production]
tdd by mattpocock — Test-driven development with a red-green-refactor loop, one vertical slice at a time. [production]
to-spec by mattpocock — Turns the current conversation into a spec and publishes it to your issue tracker. No interview, just synthesis. [production]
to-tickets by mattpocock — Breaks a plan or spec into tracer-bullet tickets, each declaring its blocking edges. [production]
triage by mattpocock — Moves issues and external PRs through a state machine of triage roles and writes agent-ready briefs. [production]
wayfinder by mattpocock — Plans work too large for one agent session as a map of decision tickets, resolved one at a time. [production]
wizard by mattpocock — Generates an interactive bash wizard for the steps only a human can do — dashboards, credentials, cutovers. [production]
writing-for-agents by mattpocock — How to write documents agents actually follow. Use it when authoring a skill or editing AGENTS.md. [production]
Hermes-native skill builders
Agent QA skills by Vostride — Three portable skills for authoring Agent QA tests, triaging failed runs from evidence, and applying scoped fixes through MCP or CLI. [beta]
hermes-skill-factory by Romanescu11 — Meta-skill that auto-generates reusable skills from your workflows. Point it at a repeated task and it creates a skill for it. [beta]
lintlang by Hermes Labs — Zero-LLM static analysis for agent configs, tool descriptions and system prompts. Catches vague tool descriptions and missing stop conditions in CI. [beta]
litprog-skill by tlehman — Literate programming skill across Claude Code, OpenCode, Hermes. Weaves code and prose into documented, executable notebooks. [beta]
maestro by ReinaMacCredy — Skill orchestration with Conductor planning + Beads tracking. Multi-step skills as observable pipelines. [beta]
model-compare by moonlight-lupin — Blind multi-model A/B comparison across tool-calling, coding and review modes, with token-efficiency metrics. [production]
pydantic-ai-skills by DougTrajano — Pydantic AI with agentskills.io. Type-safe schema validation for skill inputs/outputs. [production]
rtk-hermes by ogallotti — Compresses terminal output via RTK before it reaches LLM context. 60-90% token reduction. Zero config. [beta]
SkillClaw by AMAP-ML — Auto-evolves and dedupes your skill library from session data. Native Hermes integration. 2.5k★. [production]
skillsdotnet by PederHP — C# implementation of agentskills.io with MCP integration. [beta]
ssot-check by Conor Bronsdon — Audits repeated facts in documentation against a declared source of truth and reports stale copies. [production]
super-hermes by Cranot — Teaches Hermes to write its own analytical prompts. Meta-reasoning before execution. [experimental]
windows-terminal-hygiene by vollegrewar — Pre-flight checklist that keeps agent terminal commands from hanging on Windows: timeouts, non-interactive flags, pty, encoding, and a 3-minute recovery flow. [beta]
wondelai/skills by wondelai — Cross-platform agent skills for Claude Code and agentskills.io platforms. 2.1k★. [production]
🌐 Browser & Web
hermes-cloudflare by raulvidis — Cloudflare browser rendering plugin. Headless browsing through Cloudflare's infrastructure. [experimental]
vessel-browser by unmodeled-tyler — AI-native Linux browser with MCP control and autonomous browsing. Built for agent use, not a headless wrapper. [experimental]
✉️ Communication & Social
clawsocial-hermes-plugin by mrpeter2025 — Social discovery network. Semantic interest matching, real-time WebSocket messaging, shareable profile cards. Bilingual EN+CN. [beta]
famulor-skill by Famulor — Operates a Famulor voice-agent workspace over a hosted OAuth MCP server: assistants, call, messaging and email history, campaigns and telephony. [beta]
hermes-tag by DanielLi202 — Context-selection layer for Feishu/Lark and Slack group chats. Answers @-mentions from bounded per-chat evidence instead of full-history RAG, with per-chat memory isolation. [beta]
hermes-agent-telegram-checklist by johnsje183 — Creates and toggles native Telegram checklists in chats and forum topics over MTProto, behind a write allowlist. [beta]
hermes-tweet by Xquik — Native Hermes Agent X/Twitter plugin for tweet search, reply reading, user lookup, monitoring, posting, replies, DMs, and approval-gated X actions through Xquik. [beta]
microsoft-workspace-skill by Andrew-Girgis — Full Outlook/Hotmail/Microsoft 365 integration via Graph API. Email, calendar, contacts, free/busy. OAuth2 auto-refresh. Preview-before-send pattern. [beta]
tweetclaw by Xquik — OpenClaw plugin and agent skill to scrape tweets, search tweet replies, export followers, look up users, run media, monitors, webhooks, giveaway draws, and approval-gated posts through Xquik. [beta]
twitr-skills by lnvestor — Six X/Twitter skills: reads and search, bulk exports, keyword monitors with signed webhooks, publishing, and a presence routine. Pay-per-call in USDC over x402, no API key. [beta]
before-you-build by bin1874 — Product-risk review before implementation. Challenges demand, positioning, monetization, retention, trust, distribution, and feature-adoption risk before an agent starts building. Works with Hermes, OpenClaw, Claude Code, Codex, and other SKILL.md-compatible agents. [production]
grill-me by mattpocock — Get relentlessly interviewed by your agent until every branch of the decision tree is resolved. Use *every* time before making a change. Most popular skill in the pack. [production]
handoff by mattpocock — Compacts the current conversation into a handoff document another agent can pick up cold. [production]
hermes-plugins by 42-evey — Goal management, inter-agent bridge, model selection, cost control. Four plugins covering common operational needs. [beta]
onequery-cli by Wordbricks — CLI skill for safe, auditable queries for agents against approved data sources. [beta]
🎨 Creative & Media Generation
anti-ui-slop by UIZZE — MIT skill for design contracts, required UI states and hard finish gates on agent-generated UI. Works without an account; live reference search, validation and audits need the paid UIZZE MCP. [beta]
levea-ai-video-editor by brajendrak00068 — Agentic video editing: clip selection, captions, silence removal, vertical reframe, chroma key, MP4 export. [beta]
lora-concept-removal-tagging by kwanmxiii-bit — Tags LoRA training images and drops target concepts so the weights absorb them instead. Caption-dropout based. [beta]
OrkasVideoStudio by Orkas-AI — Local-first TypeScript CLI and MCP toolkit with 14 skills for agent-authored video composition, editing, generation, and plan-based assembly. [beta]
runapi-cli by RunAPI — Runs image, video, music and model API jobs from any CLI-capable agent. [beta]
typeui-hermes by Bergside — Design skills that give Hermes a consistent component vocabulary when generating UI; setup in the Hermes guide. [production]
🔧 DevOps & Deployment
aivpn by ZoniBoy00 — Per-location Docker WireGuard relay for Proton VPN with kill switch, SOCKS5 and prompt-injection-safe fetching. [beta]
evey-setup by 42-evey — One-command setup for full hermes-agent stack with free models and 29 plugins. [beta]
nika by supernovae-st — Deterministic workflow runner Hermes can delegate to. Repeatable jobs become reviewable .nika.yaml files with plan, cost and permit checks up front and a hash-chained trace after. [beta]
nix-hermes-agent by 0xrsydn — Nix package and NixOS module. Fully reproducible deployments via Nix flakes. [beta]
AgentCash by Merit-Systems — 300+ premium APIs + wallet for paying via x402 or MPP. Free USDC for trying out. [beta]
barker-mcp by Barker — Stablecoin yield skills: live APY index, risk signals, yield advisor. Remote MCP with x402 pay-per-call; discovery is free. [beta]
chainlink-agent-skills by Chainlink — Official Chainlink skills. Oracle data, CCIP, smart contract interaction. [production]
erpclaw by AvanSaber — AI-native open-source ERP and double-entry accounting you self-host and run in plain English. Invoicing, inventory, general ledger, payroll, multi-company books. [beta]
hermes-blockchain-oracle by gizdusum — Solana blockchain intelligence MCP server. On-chain analytics and wallet data. [experimental]
hermes-payguard by nativ3ai — Safe USDC and x402 payment plugin with spending limits and approval flows. [experimental]
mercury by hxsteric — Multi-chain blockchain cash flow analyzer with WebGL dashboard. On-chain forensics. [beta]
ripley-xmr-gateway by KYC-rip — Monero (XMR) blockchain gateway. Private cryptocurrency transactions from agent workflows. [experimental]
true402-token-safety by true402 — Pre-trade rug/honeypot check for Base tokens. Runs a real on-chain buy/sell simulation to prove a token can be sold, plus liquidity and contract structure. Free daily checks, no API key; pays per call over x402 after. [beta]
🤖 Multi-Agent & Swarms
hermes-conductor by forcewake — Kanban-orchestrated multi-harness coding swarm: controller/worker/reviewer skills with evidence-only task completion. [beta]
hermes-council by Ridwannurudeen — Adversarial multi-perspective council MCP. Multiple AI viewpoints debate before commit. [experimental]
NemoHermes by Hmbown — NVIDIA capability registry and Spark-aware routing. Routes compute-heavy tasks to GPU infrastructure. [experimental]
agent-android by AIVane Labs — LAN-first Android control over WiFi. No USB/ADB/root needed. Health checks, taps, swipes, screenshots, inspect→act→smoke flows. [beta]
argo by Allendior — Signed typed jobs from Hermes to a paired Android phone. No on-device LLM; fail-closed taps, Home, Back, and screen reads. [beta]
hermes-agent-camel by nativ3ai — Hermes with integrated CaMeL trust boundaries. Formal trust verification for safety-critical deployments. [beta]
incognito-mode by 幻灭文学出版社 — Defense-in-depth incognito mode: PID-locked sandbox, shell history suppression, 10-step reverse audit with Python secure wipe, subagent inheritance protocol. [beta]
resemble-ai/detect-skill by Resemble AI — Deepfake detection: AI-generated audio/image/video/text, source tracing, watermarking, speaker ID. [beta]
Verified Memory Vault by secondbrainstarter — Plain-Markdown Obsidian vault that gives AI coding agents persistent memory, with a health checker (protocol violations, dead links, bloat) and a git pre-commit hook that refuses commits mass-deleting memory files. [beta]
🎯 Domain & Novelty
anihermes by rodmarkun — Local anime server and tracker with NL interface. Browse, track, get recommendations via conversation. [beta]
colony-skill by TheColonyAI — Collaborative intelligence platform. AI + humans post findings, complete tasks, build reputation. [beta]
Wizards-of-the-Ghosts by Hmbown — Fantasy spell-themed skill pack. cast lint instead of npm run lint. [experimental]
zillow-skills by therohitdas — Zillow property data skills for AI agents — Zestimate, listings, photos, schools, taxes, price history — via the Zillapi REST API. MIT-0, free tier. [beta]
🔌 Plugins
Hermes plugins — installable packages that add tools, commands, hooks or providers to the agent itself, rather than a single SKILL.md. Install with hermes plugins install <repo>.
babysitter by a5c-ai — Deterministic supervision loop for agentic workforces: enforces plans, retries and completion criteria on long multi-step runs. [production]
cronalytics by 8bit64k — Analytics and observability for Hermes cron jobs. A dashboard for automations you'd otherwise never see fail. [beta]
hermes-code-bridge by xuyang-liu16 — Makes Hermes the control plane for local coding agents — Codex, Claude Code, OpenCode, Gemini CLI, Kimi Code. [beta]
Hermes Connector by Corsen AI — Connects an exact Hermes profile and session to user-selected Chrome tabs through a local companion and authenticated loopback broker. [production]
hermes-curator-evolver by pingchesu — Evidence-driven skill evolution: reports, dry-run proposals, candidate search and guarded apply. [beta]
hermes-dynamic-workflows by lingjiuu — Composes multi-step workflows at runtime instead of hard-coding them into a skill. [beta]
hermes-humalike-plugin by Humalike — Makes a chat-connected Hermes read like a person rather than a bot — pacing, typing behaviour, reply timing. [beta]
hermes-live-voice by bielcarpi — Real-time voice control. Keep talking while Hermes keeps working in the background. [beta]
hermes-telemetry by nujovich — Budget enforcement plus observability. Stops runaway spend before it happens rather than reporting it after. [beta]
planning-with-files by OthmanAdi — Crash-proof file-based planning for long-running agent tasks, with session recovery after a context clear. [production]
signetai by Signet-AI — Syncs memories, shared identity files (AGENTS.md, CLAUDE.md), transcripts and secrets between agents and machines. [beta]
🧠 Agent Profiles
Installable Hermes profiles — a persona, memory and curated skill/plugin bundle that reconfigures the whole agent for one role. Install with hermes profile install <repo-url>.
donna-starter by AtlasOmnia — Opinionated starter profile with a persona, 73 curated skills and a guided first run. A good first profile. [beta]
hermes-profile-kit by NewTurn2017 — Drop-in kit that creates four isolated profiles — coder, assistant, research, community-bot — with separate config and memory. [beta]
hermes-profile-packs by Dadmin88 — Three profile packs kept as coherent teams rather than isolated prompts: Agency for professional work, Council for personal life, Academy for teaching. Each profile ships its own SOUL.md, skills and distribution manifest. [beta]
hermes-profiles by magnus919 — Curated profiles for specialist swarms. The largest profile collection in the ecosystem. [production]
violin by Strategic-Automation — Supervised pentest profile: 31 playbooks across the OWASP, API and LLM Top 10, with interactive scoping and approval gates before any target-touching command runs. Hermes-native, no extra API keys. [production]
wisdomforge-kids-Hermes-profiles by smfworks — Parent-operated kit for standing up a private, child-facing profile in one of three age bands, with a separate profile per child rather than a shared adult one. Says plainly that it does not make AI safe; it gives a parent a design they can inspect and refuse. [beta]
💾 Memory Providers
Swap-in backends for Hermes's memory layer. A provider changes what the agent remembers and how it recalls it — install one and the built-in memory tools route through it.
agentcairn by ccf — Long-term cross-project memory backed by your own Obsidian vault. Daemonless, plain files, no opaque store. [beta]
causal-memory by JingxuanC — Memory provider with a causal core: flat facts plus typed decision→outcome causal edges (caused / enabled / prevented) on one local SQLite store. Survives c