The most popular skills and skill registries in the Hermes ecosystem, ranked by GitHub stars. Skills are reusable capabilities that follow the agentskills.io open standard — from cybersecurity to diagram generation to creative ideation.
This page ranks the whole category by stars. For picks by what you're trying to do — with install commands, compatibility notes, and a verdict per skill — see the Hermes Skills Hub.
Superpowers pairs composable skills with instructions that guide agents through a software development workflow from specification and design sign-off to planning and subagent execution. Its documented support for multiple harnesses makes it a fit for teams that want a structured development methodology across agent setups.
Open Design is a local-first system for design and prototyping, with skills, templates, design systems, and plugins. It supports several coding agents, including Hermes Agent, and also offers a native DeepSeek Harness runtime.
This repository provides a structured cybersecurity skills collection spanning 34 security domains. Its skills are mapped to frameworks including MITRE ATT&CK, ATLAS, D3FEND, and F3, and the project lists compatibility with more than 26 AI platforms.
This skill turns natural-language prompts and real sources into editable draw.io files. Its incremental sync is designed to preserve manual layout and annotations, and diagrams can be generated through the diagramctl CLI or an optional MCP server.
Internet Court provides a master router and connector skills for agent-to-agent commerce, alongside vendored copies of official protocol skills. It adds a verification and dispute-resolution layer to the skills it brings together.
Avoid AI Writing can detect writing patterns, rewrite text, or edit prose in place, with optional casual, professional, technical, warm, and blunt profiles. Its command-line detector accepts files or piped text, making it useful for both one-off checks and text-processing workflows.
SkillClaw uses background evolution to deduplicate and improve skills, and can unify skill libraries across agents, devices, and team members. It integrates with Hermes, Claude Code, Codex, and OpenAI-compatible APIs.
This cross-platform skills library includes 51 expert frameworks and 14 guided metaskills, with compatibility for Claude Code, Codex, and agent skills specifications. It also maintains session state in the project's docs folder.
This collection offers professional agent skills as plain Markdown files, with listed tasks ranging from writing PRDs to decoding leases. An installer copies the skill files, and uninstalling removes them; the README says they run natively in Hermes Agent and Claude Code.
These skills cover YouTube transcripts, video search, channel browsing, and playlist extraction, with support for Hermes Agent, OpenClaw, Claude Code, and Cursor. Use requires a TranscriptAPI key, which offers a 100-credit free tier.
Hermes Skill Factory watches active workflows to detect and propose reusable skills, then generates SKILL.md documentation and Python plugin scaffolds. Slash commands let users manage, list, and save the detected patterns.
Super Hermes teaches Hermes Agent to generate custom thinking instructions for specific tasks, with pre-built analytical lenses for various domains. It reports findings alongside a report of analysis blind spots.
Adds type-safe agentskills.io support for Pydantic AI, including remote skill registries from Git and S3. It also provides tools for bundled resources and scripts, with sandboxed execution to isolate untrusted scripts from the host.
Turns rough ideas into structured design documents and implementation handoffs for Hermes Agent. Lite and Full modes accommodate different project complexity, while a review gate evaluates specs with PASS/FAIL models.
Creates .lit.md documents that combine prose, diagrams, and source code, then extracts runnable files with a tangler. It also syncs edits to source files back into the literate document, keeping code and documentation connected.
Provides a harness for long-running agents, with SQLite storage and event logging through a mechanism kernel. It also coordinates agent teams through the SLP cooperative-agent protocol and distributes prompt-first working methods as Markdown recipes and skills.
Shares memory across instances through Markdown files and automatically captures high-value decisions and outcomes. It also includes a pipeline for fine-tuning and switching to replacement models.
A Portuguese-language Hermes Agent community project offering skills in SKILL.md format alongside installation, configuration, and usage guides. Contributions are made through forks and pull requests.
Provides prompting guidance for FLUX image and video generation, plus BFL API integration for polling and webhooks. Its FLUX 3 video suite includes eight specialist skills.
Lists nine Chainlink skills that follow the Agent Skills spec, with installation available through npx skills add smartcontractkit/chainlink-agent-skills. The README describes the repository as containing community examples only.
Creates agent instructions from repositories, documentation, and developer discourse, with both automated and step-by-step workflows. It also includes skill testing and export.
Nika is a workflow language for AI that checks workflows before calling a model and runs with a model of your choice. It also verifies a tamper-evident record for each run.
Offers a registry of reusable AI skills based on the AgentSkills specification, with a Python client and CLI. Skills can be installed with recursive dependency resolution, and the project includes commerce features using TON payments.
Supports autonomous execution of long-running, multi-step tasks with real-time tracking of progress and decisions. Built-in rollback procedures provide a way to recover safely.
This Hermes Agent skill manages Nextcloud files, notes, calendars, tasks, and contacts through WebDAV and CalDAV. It uses HTTPS with Nextcloud App Passwords and includes a guided setup script for validating credentials and configuring access.
HermesHub is a community hub for browsing, searching, and installing skills, with federated search across GitHub and Hugging Face. It implements the Agentic Resource Discovery (ARD) v0.9 specification and is described as operating autonomously around the clock via Hermes Agent on NVIDIA infrastructure.
Eagle Eye narrows the skills presented before LLM calls using a five-layer retrieval system that includes hard triggers, FTS5, synonyms, dense embeddings, and RRF fusion. It is designed to reduce token usage and can degrade gracefully when embedding models or dependencies are unavailable.
Agentic-MCP-Skill is a progressive MCP client that incrementally discovers server metadata and tool schemas through three-layer lazy loading to minimize token usage. Its socket-based daemon maintains persistent connections, making it a distinct fit for Hermes Agent skills that need to manage MCP discovery efficiently.
Life_OS is an ADHD-friendly personal execution system for habits, study blocks, and agent control through MCP. It stores data locally in a single SQLite file and includes a React web app and an Expo-based Android application.
Hermaguard performs read-only, adversarial code reviews using parallel subagents focused on edge cases, adversarial review, and integration analysis. It combines a deterministic pre-scan with proof-of-concept execution in a hardened sandbox to verify findings; it identifies problems but does not fix them.
This Hermes Agent skill coordinates Hygiene, Clarity, and Correctness agents to simplify code, supporting TypeScript, JavaScript, Python, Go, and Rust. Its tiered application process includes an approval gate, distinguishing it from a fully automatic rewrite.
Fablize for Hermes adapts verified procedures for breaking tasks into multiple stories with mandatory evidence-based verification gates. It also provides an investigation protocol for reproduction and causal debugging, plus a Python goal engine for persistent task state.
Hermes Cybersec Lab packages cybersecurity skills and tools, with highlights citing 2,077 total skills across repositories, 131+ tools for network, forensics, and exploitation, and 28 frameworks and standards. Its listed integrations include MITRE ATT&CK, MISP, CVSS, EPSS, OWASP, NIST, Sigma, and YARA.
Skillsdotnet is a C#/.NET implementation of agentskills.io with MCP integration. It uses SKILL.md files for progressive disclosure, integrates with Microsoft.Extensions.AI for tool-based skill loading, and can automatically register common agent skill directories.
Hermes HUD is a local, read-only dashboard for real-time visibility into Hermes Agent activity, including token costs, session history, skill analytics, and incidents. It belongs in a skills list as an observability companion, with optional incident alerts via Telegram and Feishu.
This collection provides plain-markdown skills for code review, session management, and related workflows, and is described as production-tested across Claude Code, Codex, Cursor, and OpenCode. Skills can be installed by copying a directory or using the bundled installer.
This Claude Code skill creates reproducible demo GIFs for a repository README using vhs or Playwright recording scripts, then optimizes them for target sizes under 8 MB. It supports demos of terminal CLIs, TUIs, web apps, and libraries.
This CLI and MCP server searches remoteopenclaw.com for more than 13,870 MCP servers and 4,384 agent skills, and provides installation commands for discovered tools and plugins. It is useful for Hermes users looking to find and install capabilities through an MCP-compatible client.
Agent Skill Builder generates Claude Code extensions and portable Agent Skills core files, then validates them against runtime-specific profiles with CI-friendly exit codes. It also checks official frontmatter specifications weekly for field changes, helping skills stay aligned with their specifications.
Skill Evolution analyzes agent session history to identify skill coverage gaps and draft structured improvement proposals. It gates updates behind deterministic, LLM-judge, and regression evaluators, and runs locally using Python’s standard library without daemons or external services.
These Hermes Agent skills route file operations using the path syntax for native Windows, Git Bash/MSYS, or WSL, and select and verify existing project-compatible Python interpreters. They address two practical tasks when using Hermes across Windows environments.
Leverage provides agent skills for social data, CRM management, and lead enrichment, with support for Claude Code, Cursor, and other agentskills.io-compliant clients. It also includes an MCP server for browser-based OAuth authentication.
This collection of Hermes skills covers Douyin video transcription into structured Markdown and SRT subtitles, Chinese AI-writing cleanup with a rule-based engine, and structured data extraction from PDFs and images using macOS Vision. Its focus is a set of concrete media and document-processing tasks.
This Hermes Agent plugin lets agents find support bounties, purchase ticket context, and submit answers for human review, with public wallet earnings records. It also lets users create a bounty board that returns a URL, API key, and USDC deposit addresses.
This skill automates USDC payments for HTTP 402 responses and supports Pay Link, x402, MPP, and AP2 protocols. It is compatible with OpenClaw, Hermes Agent, and the agentskills.io standard.
Skillseal tests agent skills for structure, functionality, and security in isolated Docker containers, then issues certificates verifiable with Ed25519 signatures. Its public API supports certificate verification and skill submission.
Scavio for Hermes Agent provides structured web data across 31 platforms through a hosted MCP server, exposing 191 tools and five ready-to-use workflow skills. The project also lists support for Claude, Cursor, Goose, and OpenHands.
This Hermes Agent skill generates structured Chinese news briefings at scales ranging from global to street level, with core conclusions and suggested actions. It uses semantic memory to incorporate user interests into search expansion, helping tailor each briefing.
This MCP server provides deterministic Western and Vedic astrology results for AI agents, with SHA-256 hashes for verification. It offers a no-auth MCP endpoint and per-call REST payments through x402 or MPP.