Hermes Agent Optional Skills Catalog: Install, Use, and Contribute

hermes-agentintermediate11 min readVerified Jul 25, 2026
Hermes Agent Optional Skills Catalog: Install, Use, and Contribute

The Hermes Agent ships with a large library of optional skills under optional-skills/ that extend the agent into domains like blockchain, MLOps, finance, creative work, security, and more. None of these skills are active by default, so you install only what you need, keeping the agent lean and focused. This catalog lists every available skill, shows you how to install and uninstall them, and explains how to contribute your own.

What it does

Optional skills are pre-built modules that give Hermes Agent new capabilities without modifying its core. Each skill lives in its own directory under optional-skills/ and comes with a dedicated documentation page covering setup, configuration, and usage. You install a skill with a single command, and it becomes available to the agent immediately. If you no longer need a skill, uninstalling it removes it cleanly.

Before you start

  • You must have Hermes Agent installed and the hermes CLI available in your shell.
  • Skills are installed from the official/ repository. No additional accounts or API keys are required for the install step itself, though individual skills may have their own prerequisites (e.g., a blockchain skill might need an RPC endpoint, a finance skill might need Excel).
  • The install command follows a strict naming convention: official/<category>/<skill>. The category and skill name must match exactly as shown in the catalog.

Installing a skill

To install any optional skill, run:

hermes skills install official/<category>/<skill>

For example, to install the Solana blockchain skill and the Flash Attention MLOps skill:

hermes skills install official/blockchain/solana
hermes skills install official/mlops/flash-attention

Each skill in the catalog below links to its full documentation page, which includes the exact install command, configuration steps, and usage examples.

Uninstalling a skill

To remove a skill you no longer need:

hermes skills uninstall <skill-name>

The <skill-name> is the short name of the skill (e.g., solana, flash-attention), not the full path.

Skill catalog

Diagram: Skill catalog

autonomous-ai-agents

SkillDescription
antigravity-cliOperate the Antigravity CLI (agy): plugins, auth, sandbox.
blackboxDelegate coding tasks to the Blackbox AI multi-model CLI.
grokDelegate coding to xAI Grok Build CLI (features, PRs).
honchoConfigure and troubleshoot Honcho memory for Hermes.
openhandsDelegate coding to OpenHands CLI (model-agnostic, LiteLLM).

blockchain

SkillDescription
evmRead-only EVM client: wallets, tokens, gas across 8 chains.
hyperliquidHyperliquid market data, account history, trade review.
solanaQuery Solana wallets, tokens, txs, and NFTs in USD.

communication

SkillDescription
one-three-one-ruleStructured decision-making framework for technical proposals and trade-off analysis. When the user faces a choice between multiple approaches (architecture decisions, tool selection, refactoring strategies, migration paths), this skill p...

creative

SkillDescription
audiocraft-audio-generationAudioCraft: MusicGen text-to-music, AudioGen text-to-sound.
baoyu-article-illustratorArticle illustrations: type × style × palette consistency.
baoyu-comicKnowledge comics (知识漫画): educational, biography, tutorial.
blender-mcpDrive Blender via the catalog blender MCP, with bpy recipes.
concept-diagramsGenerate flat, minimal educational SVG visuals as HTML.
creative-ideationGenerate ideas via named methods from creative practice.
heartmulaHeartMuLa: Suno-like song generation from lyrics + tags.
hyperframesRender MP4/WebM videos from HTML compositions.
kanban-video-orchestratorPlan and run multi-agent video production pipelines.
meme-generationCreate meme PNGs from templates with Pillow text overlay.
pixel-artPixel art w/ era palettes (NES, Game Boy, PICO-8).
tldraw-offlineDrive and script tldraw offline canvases with an agent.
unreal-mcpAutomate Unreal Engine editor scenes, actors, and renders.

data-science

SkillDescription
jupyter-notebookIterative Python via live Jupyter kernel (hamelnb).

devops

SkillDescription
inference-sh-cliRun 150+ AI apps (image, video, LLM) via inference.sh CLI.
docker-managementManage Docker containers, images, volumes, and Compose.
hermes-s6-container-supervisionModify or debug s6 services in the Hermes Docker image.
pinggy-tunnelZero-install localhost tunnels over SSH via Pinggy.
watchersPoll RSS, JSON APIs, and GitHub with watermark dedup.

dogfood

SkillDescription
adversarial-ux-testRoleplay a hostile user to find and triage UX pain points.

email

SkillDescription
agentmailGive the agent its own inbox: send and receive email.

finance

SkillDescription
3-statement-modelBuild integrated IS/BS/CF financial workbooks in Excel.
comps-analysisBuild comparable-company valuation workbooks in Excel.
dcf-modelBuild discounted cash flow valuation workbooks in Excel.
excel-authorBuild auditable financial workbooks headless via openpyxl.
lbo-modelBuild leveraged buyout workbooks with IRR/MOIC in Excel.
merger-modelBuild M&A accretion/dilution workbooks in Excel.
pptx-authorBuild PowerPoint decks headless with python-pptx.
stocksStock quotes, history, search, compare, crypto via Yahoo.

gaming

SkillDescription
minecraft-modpack-serverHost modded Minecraft servers (CurseForge, Modrinth).
pokemon-playerPlay Pokemon via headless emulator + RAM reads.

health

SkillDescription
fitness-nutritionGym workout planner and nutrition tracker. Search 690+ exercises by muscle, equipment, or category via wger. Look up macros and calories for 380,000+ foods via USDA FoodData Central. Compute BMI, TDEE, one-rep max, macro splits, and body...
neuroskill-bciConnect to a running NeuroSkill instance and incorporate the user's real-time cognitive and emotional state (focus, relaxation, mood, cognitive load, drowsiness, heart rate, HRV, sleep staging, and 40+ derived EXG scores) into responses....

mcp

SkillDescription
fastmcpBuild, test, and deploy Python MCP servers.
mcp-oauth-remote-gatewayManual OAuth for remote MCP servers on headless gateways.
mcporterList, auth, and call MCP servers/tools from the terminal.

migration

SkillDescription
openclaw-migrationImport an OpenClaw setup (memories, skills) into Hermes.

mlops

SkillDescription
huggingface-accelerateRun PyTorch training across GPUs with minimal changes.
axolotlAxolotl: YAML LLM fine-tuning (LoRA, DPO, GRPO).
chromaEmbedding database for RAG and semantic search.
clipZero-shot image classification and image-text search.
dspyDSPy: declarative LM programs, auto-optimize prompts, RAG.
faissFast vector similarity search at billion scale.
optimizing-attention-flashSpeed up long-sequence transformer training and inference.
guidanceConstrain LLM output with grammars; guarantee valid JSON.
huggingface-tokenizersFast BPE/WordPiece tokenization and custom vocab training.
instructorStructured LLM outputs validated with Pydantic.
lambda-labs-gpu-cloudOn-demand GPU cloud instances for ML training.
llavaVision-language chat: VQA, captioning, image dialogue.
modal-serverless-gpuServerless GPU cloud for ML jobs and model APIs.
nemo-curatorCurate LLM training data: dedupe, filter, PII redaction.
obliteratusOBLITERATUS: abliterate LLM refusals (diff-in-means).
outlinesOutlines: structured JSON/regex/Pydantic LLM generation.
peft-fine-tuningFine-tune large LLMs with LoRA on limited GPU memory.
pineconeManaged vector DB for production RAG and search.
pytorch-fsdpFully sharded data-parallel training for large models.
pytorch-lightningClean training loops with built-in distributed support.
qdrant-vector-searchVector search engine for production RAG systems.
sparse-autoencoder-trainingTrain sparse autoencoders to interpret model features.
segment-anything-modelSAM: zero-shot image segmentation via points, boxes, masks.
simpo-trainingReference-free preference alignment, simpler than DPO.
slime-rl-trainingRL post-training for LLMs with Megatron and SGLang.
stable-diffusion-image-generationText-to-image generation, inpainting, and img2img.
tensorrt-llmHigh-throughput LLM inference on NVIDIA GPUs.
distributed-llm-pretraining-torchtitanPretrain LLMs at scale with PyTorch 4D parallelism.
fine-tuning-with-trlTRL: SFT, DPO, GRPO, RLOO reward modeling for LLM RLHF.
unslothUnsloth: 2-5x faster LoRA/QLoRA fine-tuning, less VRAM.
whisperTranscribe and translate speech in 99 languages.

payments

SkillDescription
mpp-agentPay HTTP 402 APIs via Machine Payments Protocol (MPP).
stripe-link-cliAgent payments via Stripe Link, cards, SPT, approvals.
stripe-projectsProvision SaaS services + sync creds via Stripe Projects.

productivity

SkillDescription
canvasFetch Canvas LMS courses and assignments via API token.
here.nowPublish sites to {slug}.here.now and store files in Drives.
memento-flashcardsSpaced-repetition flashcard system. Create cards from facts or text, chat with flashcards using free-text answers graded by the agent, generate quizzes from YouTube transcripts, review due cards with adaptive scheduling, and export/impor...
shopShop catalog search, checkout, order tracking, returns.
shopifyQuery Shopify Admin/Storefront GraphQL APIs via curl.
siyuanQuery and edit a SiYuan knowledge base via its API.
telephonyProvision Twilio numbers, SMS/MMS, and AI outbound calls.

research

SkillDescription
bioinformaticsGateway to 400+ genomics and computational biology skills.
darwinian-evolverEvolve prompts/regex/SQL/code with Imbue's evolution loop.
domain-intelPassive recon of subdomains, SSL certs, WHOIS, and DNS.
drug-discoveryPharmaceutical research assistant for drug discovery workflows. Search bioactive compounds on ChEMBL, calculate drug-likeness (Lipinski Ro5, QED, TPSA, synthetic accessibility), look up drug-drug interactions via OpenFDA, interpret ADMET...
duckduckgo-searchFree keyless web, news, and image search via ddgs.
gitnexus-explorerServe an interactive codebase knowledge graph web UI.
osint-investigationFollow the money via public records and sanctions data.
parallel-cliAgent-native web search, deep research, and enrichment.
pinecone-researchAgent RAG and long-term memory with Pinecone.
qmdHybrid local search over notes, docs, and transcripts.
scraplingScrape sites with stealth browsing and Cloudflare bypass.
searxng-searchFree keyless meta-search aggregating 70+ engines.

security

SkillDescription
1passwordSet up op CLI, sign in, and read or inject secrets.
godmodeJailbreak LLMs: Parseltongue, GODMODE, ULTRAPLINIAN.
oss-forensicsSupply chain investigation, evidence recovery, and forensic analysis for GitHub repositories. Covers deleted commit recovery, force-push detection, IOC extraction, multi-source evidence collection, hypothesis formation/validation, and st...
sherlockFind accounts for a username across 400+ platforms.
unbrokerAutonomously remove your info from data-broker sites.
web-pentestAuthorized web application penetration testing, reconnaissance, vulnerability analysis, proof-based exploitation, and professional reporting. Adapts Shannon's "No Exploit, No Report" methodology with hard guardrails for scope, authoriza...

software-development

SkillDescription
code-wikiGenerate wiki docs + Mermaid diagrams for any codebase.
rest-graphql-debugDebug REST/GraphQL APIs: status codes, auth, schemas, repro.
subagent-driven-developmentExecute plans via delegate_task subagents (2-stage review).

web-development

SkillDescription
cloudflare-temporary-deployDeploy a Worker live, no account, via wrangler --temporary.
page-agentEmbed an in-page natural-language GUI copilot in web apps.

yuanbao

SkillDescription
yuanbaoYuanbao (元宝) groups: @mention users, query info/members.

Contributing Optional Skills

To add a new optional skill to the repository:

  1. Create a directory under optional-skills/<category>/<skill-name>/
  2. Add a SKILL.md with standard frontmatter (name, description, version, author)
  3. Include any supporting files in references/, templates/, or scripts/ subdirectories
  4. Submit a pull request, the skill will appear in this catalog and get its own docs page once merged

When not to use it

If you need a capability that is already built into the Hermes Agent core (e.g., basic file operations, web requests, or conversation memory), installing an optional skill is unnecessary and adds overhead. The catalog is for domain-specific or advanced functionality that would bloat the base agent. Also, if a skill's prerequisites (like a paid API key, a specific GPU, or a local service) are not met, the skill will fail silently or with confusing errors, so check each skill's documentation before installing.

Limits and gotchas

  • Skills are not sandboxed from each other. Installing many skills from different categories may cause naming conflicts or dependency clashes. Uninstall skills you no longer use.
  • The hermes skills uninstall command uses the short skill name, not the full path. If two skills share the same short name (unlikely but possible), the uninstall behavior is not documented.
  • The catalog is maintained by the community and the Hermes team. A skill's documentation page is the single source of truth for its setup and usage; the table descriptions here are summaries and may be truncated.
  • Contributing a skill requires a pull request. There is no CLI command to publish a skill directly.

What pairs with this

Browse the catalog on Neura Market

Every skill in this catalog has its own guide here, with the same metadata table and the commands reproduced from the official docs: Hermes Agent skills.

Newsletter

The #1 AI Newsletter

The most important ai updates, guides, and fixes — one weekly email.

No spam, unsubscribe anytime. Privacy policy

Related Guides