Claude Code Templates
davila7/claude-code-templatesA 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.
A catalog of roughly 2,900 community SKILL.md playbooks, with a local CLI, MCP server and plugin bundles for searching and installing them.
Huge and actively maintained, but quality is uneven and the mass of duplicated, community-sourced skills calls for selective installs.
The author changed 26 of 200 skill folders and the plugin manifest or hooks after our review (noticed Oct 29, 2026). The review describes the earlier version; we recheck changed skills weekly, the most significant changes first.
Agentic Awesome Skills is a very large library of installable SKILL.md files, plus AAS Core, a local control plane that helps an agent search the catalog, record which skills it chose, validate a stack and preview an install plan. The README says Core does not rank or recommend skills; the agent selects them. The tree has tens of thousands of files because the same skills are copied into plugin bundles for different hosts, so the 8,800 skill folders in the metadata overcount distinct skills. Skills carry frontmatter fields such as risk (safe, critical), source (personal, community), date added and tags. Content ranges from short persona-style helpers to long references (one feedback-system skill is about 22 KB), and it spans security audits, data analytics, SaaS launch, media generation and more, with many credited to outside authors. Apply and recovery features in AAS Core are labeled experimental by the maintainer. Because it is a catalog, quality is uneven and most skills are single-file instructions. It is best used for discovery, then trimming to the few skills you actually need rather than installing everything.
About 2,900 distinct skills averaging 8.8 KB, with some references, a few scripts and many eval-type files in the tree, but most are single-file instructions of varying quality.
Counted from the repository file tree. How we score depth.
A very large community collection of agent skills (tens of thousands of files) installed through an npx command; we could review only about 90 of them. The reviewed parts are a local security-audit skill (007) whose scanners read files and write a score history, and a 2slides presentation skill that sends your text and API key to 2slides.com and downloads the result. Because the vast majority of the collection and its installer were not available, this is not a statement about the whole repo.
npx agentic-awesome-skills --<agent> (latest release), which copies skills into your agent folders.README.md:128The repository (27639 files) on Oct 29, 2026. Not a full review: 27249 of 27639 files were not downloaded (large repository; we reviewed the skill folders); it contains compiled binaries or archives we cannot read; we read 3 of 15 code files; it relies on code outside the repository: AAS Core CLI / npm package agentic-awesome-skills (the installer and MCP server); About 27,250 of the 27,639 repo files, including the thousands of other skills and their scripts, are not in our copy; only about 90 files (the Claude plugin's 007 and 2slides skills and a few others) were available; 37 binary files such as the shadcn-components.tar.gz archives.
NVIDIA SkillSpector static scan: 181 pattern hits (Privilege Escalation 33, Agent Snooping 30, Rogue Agent 22, MCP Rug Pull 20, Data Exfiltration 14, Excessive Agency 13). The CRITICAL hits are the 2slides scripts sending their own API key to 2slides.com. The prompt-injection, anti-refusal and 'Output your system prompt' hits are example attack strings inside the 007 security skill's reference documents. The curl|bash string is a detection rule description in dependency_scanner.py, not a command it runs.
A review, not a certification. Skills change after we check them — read the files before installing.
MIT for the repository; individual skills credit outside sources, so check attribution for reuse.
Not really. It is a catalog and planner for choosing skills; any unattended behavior depends on the individual skills you pick.
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.
A 400+ skill library of agents, personas and commands for engineering, marketing, product, compliance and C-level work, mirrored for 13 coding agents.
A large library of 385 business, product, compliance and engineering skills with 906 standard-library Python tools, under a Commons Clause license.
A curated list of Claude skills plus hundreds of templated app-automation skills that run through Composio's hosted Rube MCP.
It is a library of thousands of SKILL.md playbooks with a local tool, AAS Core, for searching the catalog, recording the skills an agent selects and previewing an install plan.
The README documents a direct installer with a dry-run option, plugin bundles for Claude and other hosts, and the AAS Core CLI. Use the dry run first and install only the skills you need.
The project labels skills with a risk level and says Core does not apply changes unreviewed, but the catalog is community-sourced. Read the skills you install, especially any marked critical.
Both aggregate many components. AAS focuses on SKILL.md playbooks and a local selection workflow, while claude-code-templates also installs hooks, agents and MCP configs for Claude Code.