Skills pour l'agent Hermès

Apple / macOS skills — tools that interact with the Mac desktop Finder, native apps or system features accessibility, screenshots .

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

Apple Reminders via remindctl: add, list, complete.

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

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

Skills for spawning and orchestrating autonomous AI coding agents and multi-agent workflows — running independent agent processes, delegating tasks, and coordinating parallel workstreams.

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

Spawn and orchestrate autonomous AI coding agents (Claude Code, Codex, OpenCode, Antigravity, Claude Cowork) from CLI terminal tools.

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

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

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.

Configure, extend, or contribute to Hermes Agent.

Install, configure, secure, and manage the Hermes WebUI (nesquena/hermes-webui) — a self-hosted web frontend with 1:1 CLI parity.

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

Operate a Paperclip AI-coding orchestration instance — server lifecycle, agent hiring via API, adapter configuration, CLI commands, and Agent Stack architecture (Paperclip + Langfuse + local CLI agents).

Analyse, optimisation de CV + création de lettres de présentation PDF pour postes ciblés (marché Québec). Organisation par dossier d'offre.

Creative content generation — ASCII art, hand-drawn style diagrams, and visual design tools.

Dark-themed SVG architecture/cloud/infra diagrams as HTML. Includes Excalidraw hand-drawn JSON diagrams (absorbed from excalidraw).

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

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

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

Baoyu visual content generation: article illustrations, knowledge comics, and infographics with consistent type × style × palette systems. Absorbed from baoyu-article-illustrator, baoyu-comic, and baoyu-infographic.

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

Generate images, video, and audio via diffusion workflows.

Generate project ideas via creative constraints.

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

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

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

Manim CE animations: 3Blue1Brown math/algo videos.

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

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

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

Build creative browser demos with DOM-free text layout.

Configuration technique et workflows pour génération de vidéos animées programmatiques avec React/Remotion.

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

Songwriting craft and Suno AI music prompts.

Control TouchDesigner via twozero MCP.

50 styles graphiques + 99 règles d'ergonomie pour génération d'interfaces utilisateur.

Skills for data science workflows — interactive exploration, Jupyter notebooks, data analysis, and visualization.

Iterative Python via live Jupyter kernel (hamelnb).

Audit system for agent interventions (Hermes, Claude Code, subagents). 100% .md files, no database. Snapshots, rollback, chronological search. Tracks every action across months with ability to revert. Includes FR reference at references/audit-system-fr.md.

Execute multi-phase service deployments autonomously with safety guards, journaling, and a morning report — Phase 0 diagnostics, sequential deployment, anti-gaspillage budgets, escalation policy.

Visualise et gère la fenêtre de contexte — alertes à 40%, stratégies de compression, monitoring en temps réel.

Protocole complet : Coolify est le centre névralgique de l'infrastructure — tout le web, réseau, déploiements, certificats, backups, monitoring.

Diagnose and fix Hermes cron jobs showing "error" status — systematic workflow for listing, running manually, identifying root causes (exit codes, SSH, tokens, scripts), applying fixes, and verifying recovery. Covers no_agent and LLM-driven cron jobs across local and VPS targets.

Synchronize text files (logs, configs, state) between mx and vps-vmi2802045 via SSH — push, pull, and bidirectional merge patterns + VPS task failover (absorbed from vps-task-failover).

Access and manipulate PostgreSQL/MySQL databases running in Docker containers (including Coolify) from the same host via docker exec. Covers credential extraction, shell-escaping-safe SQL execution, schema exploration, Vikunja/SQlite/PostgreSQL specific patterns.

Systematic approach to integrating and monitoring self-hosted infrastructure components with API synchronization, DNS automation, Coolify deployment, and dashboard creation

Hermes Kanban — the complete playbook for both roles: orchestrator (decompose, route, don't execute) and worker (pitfalls, handoff shapes, retry diagnostics). The lifecycle itself is auto-injected into every worker's system prompt as KANBAN_GUIDANCE.

Periodically audit Linux system security — local and remote (VPS) — with CIS-inspired checks, automatic scoring, Google Sheets logging, and auto-corrective actions. Covers quick (vitals-only), daily (full), and weekly (deep) scan modes.

Politique Zero-Tokens MCP : tout futur MCP est invoqué via une skill Hermes, jamais chargé dans config.yaml. Zéro overhead token tant que la skill n'est pas utilisée. Liste des MCP recommandés avec leurs alternatives CLI/API et template de création.

Gestion rapide et intelligente des modèles LLM — changements de provider, bascule rapide, fallback local.

Gestion de n8n par API REST - workflows, credentials, programmation SMS/notifications. Instance sur n8n.iatuto.com

Playwright via CLI et outils navigateur Hermes natifs — scraping, tests E2E, QA automatisée. Zero MCP, zero overhead token.

Routing intelligent entre les 3 approches de sessions parallèles : Topics Telegram, Multi-Profiles, et Kanban Workers. Choisit automatiquement la meilleure approche selon la demande.

Configure SSH session persistence via tmux (+ Mosh) so work survives disconnection — tmux.conf, auto-attach, aliases, and diagnostic checklist.

Système de garde pour modifications critiques : backup avant changement, journal Google Sheets, rollback. Pour VPS avec services auto-hébergés.

Connect and diagnose remote machines via Tailscale: status, connectivity tests, SSH, Tailscale SSH, browser fallback. Covers the full workflow from identifying targets to establishing a working connection.

Configurer et gérer plusieurs bots Telegram avec routage contextuel dans Hermes — inventaire, tokens, Chat IDs, gateways multiples, mapping contexte→canal

Veille automatisée des offres d'emploi à la Sûreté du Québec via l'API ManitouLS

Systematic protocol for diagnosing an unresponsive remote Linux server — SSH down but network reachable. Multi-layer connectivity testing, chronological log reconstruction, differential diagnosis, and corrective action planning.

Patrons systemd user services pour VPS — Type=simple/oneshot/forking, vérification, unités persistantes. Tmux, Camofox et autres services système utilisateur.

Webhook subscriptions: event-driven agent runs.

Diagram creation skills for generating visual diagrams, flowcharts, architecture diagrams, and illustrations using tools like Excalidraw.

Passive domain reconnaissance using Python stdlib. Use this skill for subdomain discovery, SSL certificate inspection, WHOIS lookups, DNS records, domain availability checks, and bulk multi-domain analysis. No API keys required. Triggers on requests like "find subdomains", "check ssl cert", "whois lookup", "is this domain available", "bulk check these domains".

Skills for sending, receiving, searching, and managing email from the terminal.

Himalaya CLI: IMAP/SMTP email from terminal.

Framework d'évaluation stratégique à 7 axes — premortem, blindspot, L99 (GO/NO-GO), show logic, OODA, wargame, deepsync. Pipeline tri-agent: Stratège → Auditeur → Décideur. Utilisé avant toute décision GO/NO-GO sur projet, produit, ou initiative.

Skills for setting up, configuring, and managing game servers, modpacks, and gaming-related infrastructure.

Host modded Minecraft servers (CurseForge, Modrinth).

Play Pokemon via headless emulator + RAM reads.

Utilise Genspark Tool CLI (`gsk` v1.5.0+) comme sous-agent pour recherche web, crawl de pages, analyse de médias et génération image/vidéo/audio via API. USE WHEN: recherche web avec sources citées, crawl de page spécifique, capture screenshot, génération d'images/vidéos avec modèles IA, analyse de fichiers volumineux (PDF, docs), transcription audio. NO-TRIGGER: tâches de code (→ Claude Code), raisonnement complexe sans web (→ DeepSeek/Hermes), accès chat LLM interactif (→ genspark-chat skill).

Skills for searching, downloading, and working with GIFs and short-form animated media.

GitHub workflow skills for managing repositories, pull requests, code reviews, issues, and CI/CD pipelines using the gh CLI and git via terminal.

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

Complete GitHub workflow: authenticate, manage repos, create PRs, review code, triage issues — all via gh CLI and REST API. Includes codebase inspection with pygount (absorbed from codebase-inspection).

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

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

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

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

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

inference.sh Run 150+ AI applications in the cloud via the inference.sh https://inference.sh platform. One API key for everything — access image generation, video creation, LLMs, s…

Skill: info system Description Document de référence complet de la topologie système de bf. Utilisable par n'importe quel LLM pour comprendre l'environnement et prendre de bonnes d…

Log 3 canaux simultanés : fichier .md local + Google Sheets webhook + Vikunja API. Graceful degradation si canal down.

Skills for working with MCP (Model Context Protocol) servers, tools, and integrations. Documents the built-in native MCP client — configure servers in config.yaml for automatic tool discovery.

MCP client: connect servers, register tools (stdio/HTTP).

Convertit automatiquement un rapport Markdown interne en page HTML professionnelle hébergée sur le réseau Tailscale. Utilise pandoc (ou python-markdown en fallback) localement — zéro token consommé. Déclencher après toute génération de fichier .md interne (rapport, audit, veille, synthèse). Produit un lien HTML cliquable et écrit le fichier dans ~/rapports-html/. Accessible via http://mx.home/rapports/ et http://mx.tailaa682b.ts.net/ depuis toutes les machines Tailscale.

Skills for working with media content — YouTube transcripts, GIF search, music generation, and audio visualization.

Search/download GIFs from Tenor via curl + jq.

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

Analyze images, screenshots, and visual content using Google Gemini (google-genai Python SDK). User preference overrides Hermes built-in vision_analyze.

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

Spotify: play, search, queue, manage playlists and devices.

YouTube transcripts to summaries, threads, blogs.

Knowledge and Tools for Machine Learning Operations - tools and frameworks for training, fine-tuning, deploying, and optimizing ML/AI models

Model evaluation benchmarks, experiment tracking, data curation, tokenizers, and interpretability tools.

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

ML evaluation: W&B experiment tracking + lm-eval-harness LLM benchmarking (absorbed evaluating-llms-harness).

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

Model serving, quantization (GGUF/GPTQ), structured output, inference optimization, and model surgery tools for deploying and running LLMs.

LLM inference: llama.cpp local GGUF + vLLM high-throughput serving + HuggingFace Hub CLI (absorbed serving-llms-vllm and huggingface-hub).

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

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

Specific model architectures and tools — image segmentation (Segment Anything / SAM) and audio generation (AudioCraft / MusicGen). Additional model skills (CLIP, Stable Diffusion, Whisper, LLaVA) are available as optional skills.

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

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

Deploy, manage, and maintain local LLMs via Ollama — batch pull, resource planning, agentic model selection, inventory management.

ML research frameworks for building and optimizing AI systems with declarative programming.

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

Fine-tuning, RLHF/DPO/GRPO training, distributed training frameworks, and optimization tools for training LLMs and other models.

Vector similarity search and embedding databases for RAG, semantic search, and AI application backends.

Note taking skills, to save information, assist with research, and collab on multi-session planning and information sharing.

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

Skills migrated from an OpenClaw workspace.

Google Workspace CLI for Gmail, Calendar, Drive, Contacts, Sheets, and Docs.

Automate job search: LinkedIn/Indeed/Jobillico + govt portals (Quebec emplois), CV analysis, daily cron veille

Nettoie et adapte un texte structuré (markdown, tableaux, abréviations) pour une lecture orale fluide — entretiens, TTS, podcasts, présentation à voix haute.

Skill: pdf2rag Description Pipeline automatique sans supervision : scan d'un répertoire → conversion PDF/DOCX/PPTX → Markdown → classement dans /knowledge share/ → indexation dans …

Skills for document creation, presentations, spreadsheets, and other productivity workflows.

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

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

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

Integrate task management systems with spreadsheets for centralized tracking and automation

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

23 agents marketing spécialisés (SEO, copywriting, emails, analytics). Workflows, prompts et personas pour campagnes automation.

Edit text in existing PDFs via natural-language prompts.

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

Skills for extracting text from PDFs, scanned documents, images, and other file formats using OCR and document parsing tools.

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

Install and select animated petdex mascots for Hermes.

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

Pour chaque demande utilisateur, créer une entrée dans le Google Sheet de suivi avec statuts (planifiée, en cours, terminée, abandonnée) et code couleur

Teams meeting summaries, job replay, Graph subscriptions.

Interagir avec l'instance Vikunja (vikunja.iatuto.com) — authentification, requêtes API, gestion des tâches via CLI Docker

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

Jailbreak LLMs: Parseltongue, GODMODE, ULTRAPLINIAN.

Skills for academic research, paper discovery, literature review, domain reconnaissance, market data, content monitoring, and scientific knowledge retrieval.

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

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

Multi-source deep research — extract target → gather community/expert/evidence data → cross-reference → structured synthesis

Automatise la navigation vers Genspark.ai, sélectionne l'agent Chat IA, choisit un modèle IA (Claude Opus 4.8 par défaut), soumet une requête et retourne les résultats formatés. USE WHEN: (1) l'utilisateur mentionne Genspark explicitement, (2) DEBUG — quand une erreur technique complexe survient (échec API, config obscure, dépendance manquante), utiliser Genspark/Claude Opus 4.8 pour la résolution au lieu de brûler des tokens DeepSeek sur du raisonnement coûteux, (3) recherche nécessitant raisonnement approfondi avec accès web. NO-TRIGGER: erreur simple déjà en mémoire ou résolvable par web_search DDGS (token gratuit, plus rapide).

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

Query Google NotebookLM notebooks directly from Hermes Agent via Camofox browser automation. Uses the existing Google session from Genspark for zero-additional-auth access. Pure SKILL.md + lightweight Python library helper — no Patchright, no Playwright, no .venv.

Query Polymarket: markets, prices, orderbooks, history.

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

Applique RTCC/OODA de manière proactive et contextuelle — orchestration, délégation, planification, ambiguïté, système/devops. Ne s'active PAS pour les échanges simples.

Skills for controlling smart home devices — lights, switches, sensors, and home automation systems.

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

Skills for interacting with social platforms and social-media workflows — posting, reading, monitoring, and account operations.

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

Règles de compression, élagage de contexte et optimisation de tokens. Appliquer à chaque réponse.

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

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

Read the live Hermes desktop DOM/CSS over CDP.

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

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

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

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

Parallel 4-agent cleanup of recent code changes.

Throwaway experiments to validate an idea before build.

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

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

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

Write implementation plans: bite-sized tasks, paths, code.

Universal structure for providing guided tutorials, workflows, and step-by-step instructions with progress tracking.

Politique permanente de documentation et traçabilité pour toute modification système : 8 artefacts obligatoires par changement + pipeline log-all automatisé (absorbé de system-doc-log) + capture proactive de connaissances en session (absorbé de guide-capture).

Monitors task execution for skill improvement opportunities. Use this skill during ANY multi-step task, agentic workflow, or substantive work session where the agent is using tools and producing deliverables. It captures patterns, user corrections, workflow insights, and methodology worth preserving as reusable skills. Also triggers during post-task feedback discussions and when the user explicitly mentions skill observations, improvements, the observation log, skill taxonomy, or asks the agent to watch for skill opportunities. Also known as "One Skill to Rule Them All" — trigger on this phrase too. IMPORTANT: this skill MAY be invoked at the start of a task-oriented session ONLY when no other observer skill is active AND the session uses tools directly (not pure routing/delegation). In the Prisma ecosystem, prefer explicit invocation over automatic: description-level auto-trigger is disabled for OpenClaw-sourced requests. For reliable activation, pair this description with a CLAUDE.md instruction or harness-level session-start hook (see Recommended Activation Setup) — description-level matching alone is not enforceable.

Skill: tts clean Description Nettoyage de texte journalistique pour production audio TTS / podcast . Transforme un texte brut rapport, article, veille en script oral naturel, lisib…

Skill: v3ille critique Description Framework d'analyse critique des tendances IA, complémentaire au rapport de veille quotidien. Ce skill est conçu pour produire une analyse argume…

Protocole d'automatisation web responsable — éviter tout blocage/bannissement (Kijiji, LesPAC, eBay, Google). Basé sur les incidents réels 2026-08-07 (cooldown Kijiji, ERR_BLOCKED_BY_RESPONSE LesPAC).

Design System Reference & HTMX Component Library cloned from cieloverde.ca with multi-theme color variations support.

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