nexu-io / open-design
A local-first design system that generates HTML prototypes, motion graphics, and PPTX decks. It integrates with coding agents like Hermes to bridge the gap between design and implementation.
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.
A local-first design system that generates HTML prototypes, motion graphics, and PPTX decks. It integrates with coding agents like Hermes to bridge the gap between design and implementation.
A comprehensive library of 817 production-grade cybersecurity skills mapped to major frameworks like MITRE ATT&CK and NIST CSF 2.0. It provides Hermes Agent with structured expertise across 29 security domains.
Converts natural language and Mermaid text into native .drawio XML diagrams. It is ideal for visualizing codebase structures, class hierarchies, and cloud infrastructure manifests.
An auditing tool that eliminates detectable AI writing patterns through rewrite, detect, and edit modes. It uses a tiered word replacement table to ensure content feels human-authored.
An agentic evolver that enables skills to evolve collectively through automatic deduplication and quality improvements. It creates a shared evolution loop for distributing optimized skills across teams.
A cross-platform library offering specialized skill collections for UX design and sales influence. These are easily installable via CLI or the Claude plugin marketplace.
A meta-skill plugin that monitors active workflows to automatically propose and generate reusable skills. It produces both Python plugin scaffolds and SKILL.md documentation.
Provides type-safe agentskills.io support for Pydantic AI. It utilizes progressive disclosure of metadata to reduce token consumption during skill loading.
Enables Hermes Agent to extract transcripts, search videos, and browse playlists without relying on yt-dlp. It provides precise timestamp and caption extraction for video analysis.
Teaches Hermes Agent to write its own analytical prompts using dynamic 'prisms' for resilience and optimization. It includes constraint reporting to identify analysis blind spots.
A structured workflow that transforms rough ideas into formal design documents and implementation handoffs. It features Lite and Full modes with a dedicated PASS/FAIL review gate.
A Rust-based harness for long-running agents featuring structured DAG-based task management and SQLite storage. It uses explicit claim and verification gates to ensure execution accuracy.
Implements literate programming by weaving code and documentation into .lit.md files. It includes a tangler to extract runnable source code and automatically syncs edits between the two.
Provides cross-instance shared memory via markdown files to capture high-value decisions. It includes a pipeline for fine-tuning and transitioning to replacement models.
Integrates Hermes Agent with the Chainlink Oracle network for smart contract interactions. It supports CCIP, VRF, and Data Feeds via the agentskills.io specification.
The official integration for FLUX image generation, providing standardized prompting guidelines and API patterns. It supports precise hex color specification and JSON structured prompting.
Converts repositories and developer discourse into agentskills.io-compliant skills. It maps instructions to specific file lines and pinned commits for high provenance.
A registry for reusable AI skills featuring an OpenAPI-driven SDK and CLI. It supports recursive dependency resolution and integrated payment flows via Stripe and TON.
Facilitates the autonomous execution of complex, multi-step tasks with real-time progress tracking. It includes built-in rollback procedures for safe recovery from failures.
A progressive MCP client that uses three-layer lazy loading to minimize token usage. It features a socket-based daemon for persistent server management and hot reloading.
Allows Hermes Agent to manage Nextcloud files, calendars, and contacts via WebDAV and CalDAV. It ensures secure communication using HTTPS and Nextcloud App Passwords.
Rewrites and scores prompts across various surfaces like Discord and Telegram based on model reasoning capabilities. It tracks token savings and quality scores in a local SQLite database.
An adversarial code review tool that uses three parallel sub-agents to hunt for bugs and exploits. It integrates Semgrep and Bandit for deterministic pre-scanning without modifying the source code.
A C# / .NET implementation of the agentskills.io specification. It integrates with Microsoft.Extensions.AI for tool-based skill loading and progressive disclosure.
A five-layer retrieval system that uses FTS5, synonyms, and dense embeddings to find the most relevant skills. This narrows skill lists before LLM calls to significantly reduce token usage.
A multi-agent swarm that simplifies code by targeting hygiene, clarity, and correctness. It uses a risk-tiered triage system to prevent dangerous modifications.
Adapts verified procedures for grounding and evidence-based verification. It uses a Python goal engine to manage persistent task state and prevent early-stop errors.
A community hub for browsing and one-click installation of skills based on the ARD v0.9 specification. It integrates federated search across GitHub and Hugging Face.
A CLI and MCP server that allows agents to search over 13,000 MCP servers and 4,000 agent skills. It provides direct installation commands for discovered tools.
Automates payments for HTTP 402 responses using USDC. It supports multiple protocols including Pay Link and x402 for seamless agent-to-agent transactions.