HarnessKit

More than a skill manager β€” manage skills, MCP servers, plugins, hooks, CLIs, configs, memory & rules across every AI coding agent. 🌟 Star if you like it!

What is HarnessKit?

More than a skill manager β€” manage skills, MCP servers, plugins, hooks, CLIs, configs, memory & rules across every AI coding agent. 🌟 Star if you like it!

Overview of HarnessKit

What it is

HarnessKit is a free, open-source application that manages skills, MCP servers, plugins, hooks, CLIs, configs, memory, and rules across multiple AI coding agents. It works on desktop, CLI, or web and brings all agent-related files under one unified interface.

Key features

  • Full Suite Extension Management – Manages five extension types (Skills, MCP Servers, Plugins, Hooks, Agent-first CLIs) across 17 agents. Supports smart organization by type, agent, or source, automatic grouping into packs, full visibility of permissions and trust scores, and one-click cross-agent deployment with automatic format conversion.
  • Agent Configs, Memory & Rules – Manages config files, memory, rules, subagents, and ignore files for 17 agents. Automatically discovers global and per-project configs, provides per-agent dashboards with live preview, supports custom paths, and reflects modifications in real time.
  • Security Audit & Permission Transparency – Built-in security engine with 18 static analysis rules assigns a Trust Score (0–100) to each extension, grouped into Safe (80+), Low Risk (60–79), and Needs Review (below 60). One-click full audit, precise file-and-line tracing, per-agent independent scanning, and permission transparency across filesystem, network, shell, database, and environment variables.
  • Marketplace Ecosystem – Three integrated marketplaces: Skills (skills.sh registry, Git URL, or local directory), MCP Servers (Smithery registry), and Agent-first CLIs. Each listing shows description, install count, and source. Skills include documentation preview and third-party security audit scores.
  • Kits – Pack skills, MCP servers, rules, and memory into portable .hk-kit.zip bundles. Deploy a Kit to any project with one click, track deployment locations, and cleanly remove. Origin tracking merges Kit-installed extensions with marketplace sources.
  • Project-Level Management – Sidebar scope picker switches between Global, All scopes, or any registered project. Agents, Extensions, and Audit filter by the active scope, allowing independent per-project setups.
  • In-Place Management – Works directly with agents’ native directories without copying files. No shadow copies or sync conflicts; reads and writes to each agent’s own config directory.

Use cases

  • Unify management of extensions, configs, memory, and rules across multiple AI coding agents (e.g., Claude Code, Codex, Gemini CLI, Cursor, Copilot).
  • Deploy a skill, MCP server, or plugin to any agent that lacks it with one click, handling format differences automatically.
  • Audit all installed extensions for security vulnerabilities and view permission details before deciding to keep or remove them.
  • Create portable Kits containing a curated set of extensions and rules, then deploy the same Kit to multiple projects.
  • Switch between global and project-specific scopes to manage agent configurations independently per project.

Getting started

See the repository README for installation and quick-start instructions.

License and pricing

Licensed under Apache-2.0. Open source and free.

Based on the project's public README and repository metadata.

Add this badge to your README

Featured on YunoDB Badge styles & help

Maintain this project? Show that it is featured on YunoDB.

Markdown[![Featured on YunoDB](https://yunodb.com/badge/realzstharnesskit.svg)](https://yunodb.com/tools/realzstharnesskit?utm_source=github&utm_medium=badge)
HTML<a href="https://yunodb.com/tools/realzstharnesskit?utm_source=github&utm_medium=badge"><img src="https://yunodb.com/badge/realzstharnesskit.svg" alt="Featured on YunoDB" height="20"></a>