Magnus919 Agent Skills
A catalog of nearly two hundred skills: code delivery, review, debugging, AI governance, product work and agent coordination.
- 175 ★ on the repository
- Last commit
- MIT
Minimal access
Why this level
- The skills contain procedures and guidance and installation goes through the Skills CLI into the agent skills directory.
Install
- Works in
In your terminal, with SkillFoxx CLI
npx skillfoxx add skills/magnus919-agent-skillsDetects the agents on your machine, shows the access level and pins the version.
Other ways to install
Access and checks
How we reviewMinimal accessThe tool runs nothing and makes no network calls.
- The skills contain procedures and guidance and installation goes through the Skills CLI into the agent skills directory.
This is third-party code. Review the repository files before installing.
What it does
The repository collects more than a hundred and eighty skills in the open Agent Skills format: procedures, protocols and knowledge packs for agentic systems. A "Find a workflow" table maps a job to a starting skill: `neckbeard` takes an issue through implementation to a verified pull request, `implementation-planning` plans execution of an approved spec, `code-review` reviews a patch with evidence, `systematic-debugging` finds a failure cause before repair, `verified-delivery` carries an authorized change through merge and `harness-engineering` fixes an unreliable agent and builds its runtime. Separate groups cover product (`product-discovery`, `product-strategy`), operations and executive support, AI governance (`ai-governance`, `ai-operating-economics`) and coordinating a team of agents (`multi-agent-coordination`). Each skill states its boundary. Review assesses a patch and writes no fix, and delivery grants no merge rights. It installs through the Skills CLI on Codex, Claude Code, OpenCode, GitHub Copilot, Cursor, Gemini CLI, OpenClaw and Hermes Agent.
Who it is for. Engineering leads and people building processes around agents who want ready-made responsibility boundaries between roles.
Good fit when
- An issue from the tracker has to reach a verified pull request through a clear lifecycle.
- A patch or pull request needs a review grounded in evidence, with no fixing on its own.
- An agent behaves unreliably and instructions, tools, state and recovery need rethinking.
- A decision is being made whether an agent may act and whether a new capability may enter the environment.
Not a fit when
- You need one deep skill for a narrow task: the catalog is broad and some skills are overviews.
- You want installation without Node.js and Git: the Skills CLI needs Node.js 22.20 or newer and Git.
- You want ready commands for a specific stack such as Next.js or Laravel: the skills describe processes.
Example request
Use the `systematic-debugging` skill: find the cause of the failing test in `tests/payments` and describe the diagnosis before any fix.Limitations
Installation goes through the Skills CLI and needs Node.js 22.20 or newer with npm and Git. The full catalog installs with a `--skill '*'` command and the quotes around the asterisk are required. Hermes needs the `--global` flag because its project skills folder is not auto-discovered. Updates overwrite installed files, so keep local edits separate and remove skills deleted upstream by hand. Native plugins and mounted checkouts have separate instructions in `docs/harness-installation.md`.
How to disable. Remove the skills with your installer's removal command or delete the folders from the agent skills directory. Choose one install method per agent to avoid duplicates.
FAQ
How do I install the whole catalog?
For Claude Code: `npx skills add magnus919/agent-skills --skill '*' --agent claude-code --global --yes`. For Codex change the agent to `codex`, for Cursor to `cursor`. The quotes around `*` make the shell pass the character to the CLI.
How do I pick a single skill?
Look at the list with `npx skills add magnus919/agent-skills --list` and install a name with `--skill`, for example `--skill systematic-debugging`. The name comes from `SKILL.md`.
Which agents does the catalog support?
Codex, Claude Code, OpenCode, GitHub Copilot, Cursor, Gemini CLI, OpenClaw and Hermes Agent through the Skills CLI. A guide covers other environments and known limits.
How do I update installed skills?
Rerun the full-catalog install weekly to get new skills too. Only installed ones update with `npx skills update --global` or `--project` from the project folder.
Does the verified-delivery skill grant merge rights?
No. It carries an already authorized change through merge and verification but grants no permission to merge or deploy.
Discussion
No questions or reviews yet.
Related
Skills90 specialist skills and the Nexus orchestrator for Claude Code, Codex and Antigravity: bugs, tests, security, review, UX
A command-line skill manager: installs skills from a marketplace, GitHub, private Git and npm into 45 agents' folders and scans them first
A shared registry of 790+ skills for Claude Code, Gemini, Cursor, Codex and five more agents, installed as a set or one by name.
A skill registry with a CLI and API: the agent describes a task and gets a recommendation with a trust score and install command
A collection of 14 portable skills for long agent runs: work planning, live app QA, search across past sessions and a phone call
A base AGENTS.md and 12 skills: delegation to Codex, Grok, Devin and Antigravity, planning, copywriting, scoped changes.
Comparison with similar tools
Cards that share the main search query, and how they differ.
| Tool | Official | Repository stars | Main difference |
|---|---|---|---|
| Magnus919 Agent Skills (this card) | no | 175 | |
| Superpowers | no | 316,679 | a development methodology in one set; the Magnus919 catalog adds product, operations and AI governance skills with explicit boundaries |
| Tech Leads Club Agent Skills | no | 7,161 | another skill catalog; this one has job routes and boundaries between skills |
README in short
The catalog says where to start by job and keeps boundaries: each skill states what it does not do and who owns that. Detailed trust routes live in `references/agent-trust-routes.md`.