A large catalog and npx installer (aitmpl.com) of agents, commands, hooks, MCP configs and skills for Claude Code, much of it collected from other authors.
Handy discovery tool with a huge catalog, but it is an aggregator of uneven quality and its installer pulls components from a remote catalog at run time.
Claude Code Templates is a command-line installer and web catalog rather than one skill. Running npx claude-code-templates@latest lets you browse and add components to a project: agents, slash commands, settings, hooks, MCP server configs and skills. The repo holds close to 10,000 files and about 900 skill folders, which come from the maintainer, from community contributors and from imported third-party collections; the README's credits section names sources such as K-Dense-AI's scientific skills, obra/superpowers, alirezarezvani/claude-skills and wshobson/agents, each under its original license. Skill quality therefore varies. The ones we sampled range from a 30 KB OWASP security reference covering six standards, to short wrappers for the AI Maestro agent-management CLI. Skill folders average about 10 KB, with few scripts and a large number of eval-type files in the tree. The CLI also has an analytics view (npx claude-code-templates@latest --analytics) for monitoring Claude Code usage. Think of it as a shopping aisle: useful for discovery and one-command setup, but each component should be read before you enable it.
About 917 skill folders averaging 10 KB, mostly instruction-only with a handful of references and scripts across 15 sampled folders; breadth rather than depth.
Counted from the repository file tree. How we score depth.
This skill asks for more access than most, or sends data to another service. That is usually a design choice rather than bad intent, but you should know before you install it.
Bash (plus Read, Write, WebFetch) for a skill that scrapes GitHub/HN pages and runs a user-defined SEND_COMMAND to email a daily digest.cli-tool/components/skills/ai-research/ai-release-hunter/SKILL.md:5A huge catalog of agent skills, agents, commands, hooks and MCP configs installed through the npx claude-code-templates@latest CLI. We hold under 3% of the repo and none of the installer's source, so this is a partial review. In what we hold, the OWASP skill's example code is intentionally vulnerable teaching code, and one skill (ai-release-hunter) pre-approves unrestricted Bash and emails a digest through a command you configure.
npx claude-code-templates@latest ..., which downloads and runs the newest installer that writes agents, commands, settings, hooks and MCP configs into your project/Claude config; its code is not in our snapshot.README.md:74The repository (9795 files) on Oct 29, 2026. Not a full review: 9553 of 9795 files were not downloaded (large repository; we reviewed the skill folders); it contains compiled binaries or archives we cannot read; we read 0 of 6 code files; it relies on code outside the repository: 9,553 of 9,795 repo files are not in our snapshot, including the claude-code-templates CLI source (cli-tool/src), agents, commands, hooks, settings, MCP configs and most skills; npm package claude-code-templates (installer and its analytics/chat-tunnel features); 3 binary archives (shadcn-components.tar.gz in artifacts-builder / web-artifacts-builder scripts).
NVIDIA SkillSpector static scan: 157 pattern hits (MCP Rug Pull 85, Excessive Agency 19, Privilege Escalation 18, Prompt Injection 7, Tool Misuse 6, Supply Chain 6). Most SkillSpector hits are in the OWASP security skill, whose Markdown and example files quote attack strings ('Ignore previous instructions', privileged: true, /etc/passwd, .env) as things to detect; 'concealed executable' hits are those example files; the rest are unpinned npx mentions and documentation snippets in AI/ML framework skills.
A review, not a certification. Skills change after we check them — read the files before installing.
MIT for the repository; imported components keep their original licenses, as the README's credits section states.
Not really. It is a catalog and installer; whether a component can run unattended depends on which one you pick.
A 400+ skill library of agents, personas and commands for engineering, marketing, product, compliance and C-level work, mirrored for 13 coding agents.
A curated list of Claude skills plus hundreds of templated app-automation skills that run through Composio's hosted Rube MCP.
A multi-harness plugin marketplace with 94 plugins, 202 agents, 185 skills and 105 commands for Claude Code, Codex, Cursor, OpenCode, Copilot, Antigravity and Pi.
A catalog of roughly 2,900 community SKILL.md playbooks, with a local CLI, MCP server and plugin bundles for searching and installing them.
It is a CLI and website (aitmpl.com) for installing ready-made Claude Code components: agents, commands, settings, hooks, MCP integrations and skills. It also has an analytics command for monitoring Claude Code.
Run npx claude-code-templates@latest in your project and choose components from the interactive menu. The README also documents flags for installing specific items.
We found no hidden code in the excerpts we read, but it installs hooks and MCP configs that run on your machine, and much of the catalog comes from other authors. Treat each component like any third-party plugin and read it first.
A curated list only links to projects. Claude Code Templates installs components for you, which is faster but means you are trusting its catalog and its npm package.