NousResearch/hermes-agent

4-phase root cause debugging: understand bugs before fixing.

TDD: enforce RED-GREEN-REFACTOR, tests before code.

Parallel 4-agent cleanup of recent code changes.

Throwaway experiments to validate an idea before build.

Pre-commit review: security scan, quality gates, auto-fix.

Debug Python: pdb REPL + debugpy remote (DAP).

Write a markdown plan to .hermes/plans/; no execution.

Author in-repo SKILL.md files: frontmatter and structure.

Debug Node.js via --inspect + Chrome DevTools Protocol CLI.

Exploratory QA of web apps: find bugs, evidence, reports.

Control Philips Hue lights, scenes, rooms via OpenHue CLI.

X/Twitter via xurl CLI: raw post search, posting, DM, media.

Write ML papers for NeurIPS/ICML/ICLR: design→submit.

Karpathy's LLM Wiki: build/query interlinked markdown KB.

Query Polymarket: markets, prices, orderbooks, history.

Monitor blogs and RSS/Atom feeds via blogwatcher-cli tool.

Search arXiv papers by keyword, author, category, or ID.

Create, read, edit Excel .xlsx spreadsheets and CSVs.

Create, read, edit .pptx decks, slides, notes, templates.

Teams meeting summaries, job replay, Graph subscriptions.

Create, merge, split, fill, and secure PDF files.

Notion API + ntn CLI: pages, databases, markdown, Workers.

Extract text from PDFs/scans (pymupdf, marker-pdf).

Edit text in existing PDFs via natural-language prompts.

Geocode, POIs, routes, timezones via OpenStreetMap/OSRM.

Create, read, edit Word .docx documents and templates.

Gmail, Calendar, Drive, Docs, Sheets via gws CLI or Python.

Read, search, create, and edit notes in the Obsidian vault.

Airtable REST API via curl. Records CRUD, filters, upserts.

HuggingFace hf CLI: search/download/upload models, datasets.

llama.cpp local GGUF inference + HF Hub model discovery.

vLLM: high-throughput LLM serving, OpenAI API, quantization.

W&B: log ML experiments, sweeps, model registry, dashboards.

lm-eval-harness: benchmark LLMs (MMLU, GSM8K, etc.).

Audio spectrograms/features (mel, chroma, MFCC) via CLI.

YouTube transcripts to summaries, threads, blogs.

Clone/create/fork repos; manage remotes, releases.

Search/download GIFs from Tenor via curl + jq.

Create, triage, label, assign GitHub issues via gh or REST.

GitHub PR lifecycle: branch, commit, open, CI, merge.

Review PRs: diffs, inline comments via gh or REST.

Inspect codebases w/ pygount: LOC, languages, ratios.

GitHub auth setup: HTTPS tokens, SSH keys, gh CLI login.

Himalaya CLI: IMAP/SMTP email from terminal.

Control TouchDesigner via twozero MCP.

Throwaway HTML mockups: 2-3 design variants to compare.

Songwriting craft and Suno AI music prompts.

Build creative browser demos with DOM-free text layout.

54 real design systems (Stripe, Linear, Vercel) as HTML/CSS.

p5.js sketches: gen art, shaders, interactive, 3D.

Humanize text: strip AI-isms and add real voice.

Manim CE animations: 3Blue1Brown math/algo videos.

Author/validate/export Google's DESIGN.md token spec files.

Hand-drawn Excalidraw JSON diagrams (arch, flow, seq).

Design one-off HTML artifacts (landing, deck, prototype).

Generate images, video, and audio via diffusion workflows.

ASCII video: convert video/audio to colored ASCII MP4/GIF.

Infographics: 21 layouts x 21 styles (信息图, 可视化).

Dark-themed SVG architecture/cloud/infra diagrams as HTML.

ASCII art: pyfiglet, cowsay, boxes, image-to-ascii.

Delegate coding to OpenCode CLI (features, PR review).

Drive the user's desktop in the background — clicking, typing, scrolling, dragging — without stealing the cursor, keyboard focus, or switching virtual desktops / Spaces. Cross-platform: macOS, Windows, Linux. Works with any tool-capable model. Load this skill whenever the `computer_use` tool is available.

Use, configure, theme, extend, and orchestrate Hermes Agent.

Delegate coding to OpenAI Codex CLI (features, PRs).

Delegate coding to Claude Code CLI (features, PRs).

Send and receive iMessages/SMS via the imsg CLI on macOS.

Apple Reminders via remindctl: add, list, complete.

Track Apple devices/AirTags via FindMy.app on macOS.

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

Join a Google Meet call, transcribe live captions, optionally speak in realtime, and do the followup work afterwards. Use when the user asks the agent to sit in on a meeting, take notes, summarize, respond in-call, or action items from it.

Manage Apple Notes via memo CLI: create, search, edit.

Embed an in-page natural-language GUI copilot in web apps.

Execute plans via delegate_task subagents (2-stage review).

Deploy a Worker live, no account, via wrangler --temporary.

Generate wiki docs + Mermaid diagrams for any codebase.

Debug REST/GraphQL APIs: status codes, auth, schemas, repro.

Autonomously remove your info from data-broker sites.

Authorized 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, authorization, and aux-client leakage. Active testing against running applications you own or have written authorization to test.

Supply 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 structured forensic reporting. Inspired by RAPTOR's 1800+ line OSS Forensics system.

Find accounts for a username across 400+ platforms.

Jailbreak LLMs: Parseltongue, GODMODE, ULTRAPLINIAN.

Set up op CLI, sign in, and read or inject secrets.

Scrape sites with stealth browsing and Cloudflare bypass.

Free keyless meta-search aggregating 70+ engines.

Agent-native web search, deep research, and enrichment.

Agent RAG and long-term memory with Pinecone.

Hybrid local search over notes, docs, and transcripts.

Follow the money via public records and sanctions data.

Pharmaceutical 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 profiles, and assist with lead optimization. Use for medicinal chemistry questions, molecule property analysis, clinical pharmacology, and open-science drug research.

Free keyless web, news, and image search via ddgs.

Serve an interactive codebase knowledge graph web UI.

Evolve prompts/regex/SQL/code with Imbue's evolution loop.

Passive recon of subdomains, SSL certs, WHOIS, and DNS.

Gateway to 400+ genomics and computational biology skills.

Query and edit a SiYuan knowledge base via its API.

Provision Twilio numbers, SMS/MMS, and AI outbound calls.

Query Shopify Admin/Storefront GraphQL APIs via curl.

Shop catalog search, checkout, order tracking, returns.

Publish sites to {slug}.here.now and store files in Drives.

Spaced-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/import decks as CSV.

Fetch Canvas LMS courses and assignments via API token.

Agent payments via Stripe Link — cards, SPT, approvals.

Provision SaaS services + sync creds via Stripe Projects.

Transcribe and translate speech in 99 languages.

Pay HTTP 402 APIs via Machine Payments Protocol (MPP).

TRL: SFT, DPO, GRPO, RLOO reward modeling for LLM RLHF.

Unsloth: 2-5x faster LoRA/QLoRA fine-tuning, less VRAM.

Pretrain LLMs at scale with PyTorch 4D parallelism.

Axolotl: YAML LLM fine-tuning (LoRA, DPO, GRPO).

Text-to-image generation, inpainting, and img2img.

High-throughput LLM inference on NVIDIA GPUs.

RL post-training for LLMs with Megatron and SGLang.

Train sparse autoencoders to interpret model features.

Reference-free preference alignment, simpler than DPO.

DSPy: declarative LM programs, auto-optimize prompts, RAG.

Clean training loops with built-in distributed support.

Vector search engine for production RAG systems.

Fully sharded data-parallel training for large models.

Managed vector DB for production RAG and search.

Fine-tune large LLMs with LoRA on limited GPU memory.

Curate LLM training data: dedupe, filter, PII redaction.

OBLITERATUS: abliterate LLM refusals (diff-in-means).

SAM: zero-shot image segmentation via points, boxes, masks.

Vision-language chat: VQA, captioning, image dialogue.

Serverless GPU cloud for ML jobs and model APIs.

On-demand GPU cloud instances for ML training.

Structured LLM outputs validated with Pydantic.

Outlines: structured JSON/regex/Pydantic LLM generation.

Constrain LLM output with grammars; guarantee valid JSON.

Fast BPE/WordPiece tokenization and custom vocab training.

Speed up long-sequence transformer training and inference.

Zero-shot image classification and image-text search.

Fast vector similarity search at billion scale.

Run PyTorch training across GPUs with minimal changes.

Embedding database for RAG and semantic search.

Import an OpenClaw setup (memories, skills) into Hermes.

Manual OAuth for remote MCP servers on headless gateways.

List, auth, and call MCP servers/tools from the terminal.

Connect 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. Requires a BCI wearable (Muse 2/S or OpenBCI) and the NeuroSkill desktop app running locally.

Build, test, and deploy Python MCP servers.

Gym 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 fat — pure Python, no pip installs. Built for anyone chasing gains, cutting weight, or just trying to eat better.

Play Pokemon via headless emulator + RAM reads.

Stock quotes, history, search, compare, crypto via Yahoo.

Host modded Minecraft servers (CurseForge, Modrinth).

Build M&A accretion/dilution workbooks in Excel.

Build PowerPoint decks headless with python-pptx.

Build leveraged buyout workbooks with IRR/MOIC in Excel.

Build discounted cash flow valuation workbooks in Excel.

Build auditable financial workbooks headless via openpyxl.

Build comparable-company valuation workbooks in Excel.

Build integrated IS/BS/CF financial workbooks in Excel.

Roleplay a hostile user to find and triage UX pain points.

Give the agent its own inbox: send and receive email.

Zero-install localhost tunnels over SSH via Pinggy.

Poll RSS, JSON APIs, and GitHub with watermark dedup.

Modify or debug s6 services in the Hermes Docker image.

Manage Docker containers, images, volumes, and Compose.

Iterative Python via live Jupyter kernel (hamelnb).

Run 150+ AI apps (image, video, LLM) via inference.sh CLI.

Drive and script tldraw offline canvases with an agent.

Automate Unreal Engine editor scenes, actors, and renders.

Pixel art w/ era palettes (NES, Game Boy, PICO-8).

Create meme PNGs from templates with Pillow text overlay.

Plan and run multi-agent video production pipelines.

Generate ideas via named methods from creative practice.

HeartMuLa: Suno-like song generation from lyrics + tags.

Render MP4/WebM videos from HTML compositions.

Drive Blender via the catalog blender MCP, with bpy recipes.

Generate flat, minimal educational SVG visuals as HTML.

Article illustrations: type × style × palette consistency.

Knowledge comics (知识漫画): educational, biography, tutorial.

AudioCraft: MusicGen text-to-music, AudioGen text-to-sound.

Query Solana wallets, tokens, txs, and NFTs in USD.

Structured 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 produces a 1-3-1 format: one clear problem statement, three distinct options with pros/cons, and one concrete recommendation with definition of done and implementation plan. Use when the user asks for a "1-3-1", says "give me options", or needs help choosing between competing approaches.

Hyperliquid market data, account history, trade review.

Delegate coding to OpenHands CLI (model-agnostic, LiteLLM).

Read-only EVM client: wallets, tokens, gas across 8 chains.

Delegate coding to xAI Grok Build CLI (features, PRs).

Configure and troubleshoot Honcho memory for Hermes.

Operate the Antigravity CLI (agy): plugins, auth, sandbox.

Delegate coding tasks to the Blackbox AI multi-model CLI.