The 5 Best Claude Code Marketplace Alternatives in 2026

Anthropic's official plugin directory for Claude Code. Plugins are bundles that can include skills, agents, hooks, MCP servers, and other capabilities, and the marketplace lets you discover and install them without building them yourself. It's available automatically when you start Claude Code: run /plugin, use the Discover tab to browse, or install directly with /plugin install. If a plugin is missing you can refresh with /plugin marketplace update. The official repo is described as an Anthropic-managed directory of high-quality Claude Code plugins covering skills, agents, commands, hooks, and LSPs. Before plugins existed, teams shared scattered CLAUDE.md files, hook scripts, and MCP configs; the marketplace makes those pieces installable and shareable. The trade-off is curation: the official catalog is far smaller than community collections, and it only applies to Claude Code.

pricing See pricing

• + Zero-friction install: it's already inside Claude Code, no extra tooling to set up
• + Anthropic-curated, so a baseline quality bar exists
• + Covers the full plugin format: skills, agents, commands, hooks, LSPs and MCP servers
• + Refreshable in place with /plugin marketplace update

• − Small catalog compared to community collections (about 32k stars for the official repo vs 229k for Ecc)
• − No independent security scanning or benchmarking of listed plugins
• − Claude Code only; useless if your team also runs Codex, Cursor, or Gemini CLI
• − Requires a paid Claude Code plan, which starts at $20/month with no free tier

https://github.com/anthropics/claude-plugins-official

SkillzBench is AIQ Labs' free, independent directory of AI agent skills, plus a free MCP tool that connects any agent to the skills you choose. It addresses the problem nobody else on this list does: agent skills are everywhere, but published research found roughly 26% of them carry at least one vulnerability. SkillzBench continuously crawls and dedupes skills from GitHub, pins every version to a commit SHA, validates and license-checks each one, security-scans it, and quality-judges it. Every skill is then benchmarked in a sandbox with the same model, with and without the skill, and the resulting Lift is published in percentage points. A 0-100 trust score combines security, quality, license, maintenance and community signals, and money never touches scores. Users get per-agent loadouts: toggle skills and tools per agent, and the agent only sees what was chosen. It also tracks upstream changes, re-scans, re-benchmarks, and notifies you. Everything is free: the directory, the MCP connector, benchmarks and alerts.

pricing 0

• + The only option here with independent security scanning, license checks and measured Lift benchmarks
• + Works across Claude Code, Codex, Cursor, Gemini CLI and chat-app agents
• + Completely free, including the MCP connector and alerts
• + Trust score is firewalled from revenue: ad money never influences scores or listings
• + Per-agent loadouts stop the context-window bloat you get from loading every tool

• − It's a verification and benchmarking layer, not a plugin runtime: you still install or load skills into your agent
• − Newer platform, so it lacks the name recognition of Anthropic's official directory
• − Advertising around blog articles is the revenue model, which some readers may notice

https://skillzbench.com

aitmpl.com indexes 34 Claude Code plugin, skill and component marketplace collections from GitHub in one browsable directory. Each collection is listed with its author, star count, and capability tags (skills, agents, commands, hooks, LSPs, MCPs), so you can quickly see what a collection contains before visiting the repo. It surfaces the biggest names in the ecosystem: Ecc by affaan-m at 229k stars (a performance optimization system with skills, instincts, memory and security), Claude Mem at 87k for persistent context across sessions, Anthropic's official Claude Plugins at 32k, Claude Hud at 26k for context and tool monitoring, and Compound Engineering's plugin at 23k. It also tracks niche collections like HashiCorp's agent skills, Timescale's Postgres MCP, and Playwright browser automation. The directory is a discovery layer: it points you at GitHub repos rather than installing anything for you, and rankings are popularity-based rather than security-based.

pricing See pricing

• + Fastest way to see the whole ecosystem, including the largest collections like Ecc (229k stars)
• + Capability tags make it easy to find exactly the type of extension you need
• + Free to browse with no account

• − Discovery only: no install tooling, no verification, no benchmarks
• − Star counts measure popularity, which is a weak proxy for safety or quality
• − You still have to evaluate each repo's license and maintenance yourself

https://www.aitmpl.com/plugins/

A curated list of Claude Code plugins in the classic 'awesome list' format: slash commands, subagents, MCP servers, and hooks, all in one GitHub repo. It's one of the third-party marketplaces the Composio plugin guide highlights, and you can add it to Claude Code with the standard /plugin marketplace add owner/repo-name workflow. As a curated list it's lighter weight than a full marketplace: there's no package manager or runtime, just pointers to quality repos with short descriptions. That makes it a good starting point when you know roughly what capability you want (a code-review workflow, a notification hook, an LSP) and want a human-filtered shortlist rather than everything GitHub has to offer. It shows about 875 stars in the aitmpl.com directory, reflecting a solid community following rather than top-tier scale.

pricing See pricing

• + Human-curated, so signal-to-noise is better than raw GitHub search
• + Standard marketplace-add workflow, no new tooling to learn
• + Free and open on GitHub

• − Curation is community effort with no security scanning or benchmarks
• − A list, not a marketplace runtime: no version pinning or update management
• − Claude Code-focused, limited use for other agents

https://github.com/ccplugins

An open-source marketplace that goes big on volume: by its own listing it packages 425 plugins, 2,810 skills, and 200 agents for Claude Code, hosted at tonsofskills.com with its own CLI package manager called ccpi. Having a dedicated package manager sets it apart from plain awesome-lists: you get install and management workflows closer to what a real marketplace offers, rather than copy-pasting git clone commands. The collection is built by jeremylongshore and shows about 2.5k stars in the aitmpl.com directory. For small teams that want breadth, this is one of the largest single catalogs available. The flip side of that volume is the same one that applies to any large, lightly filtered catalog: with thousands of skills and no independent scanning or benchmarking, you're trusting the maintainers' curation and your own review process for anything that touches production code or sensitive data.

pricing See pricing

• + One of the largest single catalogs of Claude Code skills and plugins
• + Ships its own CLI package manager, so management is easier than raw git clones
• + Free and open-source

• − Volume without verification: no independent security scans or benchmarks
• − Self-reported counts with no third-party quality assessment
• − Claude Code-focused, so cross-agent teams get partial value

https://tonsofskills.com

A curated Claude Code plugin marketplace maintained by obra, showing about 1.1k stars in the aitmpl.com directory. It tags itself simply as a curated marketplace and lists commands, MCPs, and skills among its contents. The pitch is the opposite of tonsofskills.com: instead of thousands of entries, it's a deliberately small shelf where a maintainer has already done the filtering. That makes it a reasonable middle ground between Anthropic's official directory (very small, first-party) and the mega-collections (huge, unfiltered). It installs through the standard third-party marketplace flow, /plugin marketplace add owner/repo-name, so there's nothing new to learn. Like every community marketplace here, it offers no independent security scanning or benchmark data, so the trust you place in it is trust in the curator's judgment. It's Claude Code-focused and doesn't extend to other coding agents.

pricing See pricing

• + Human curation keeps the catalog small and browsable
• + Standard install workflow, no new tooling
• + Free and open-source

• − Small selection by design
• − No security scanning, license checks or benchmarks
• − Claude Code only

https://github.com/obra

A single hub, maintained by davepoon, for finding Claude skills, agents, commands, hooks, plugins, and marketplace collections. What distinguishes it from other lists is its scope: it indexes extensions for Claude Code, Claude Desktop, the Agent SDK, and OpenClaw, so it's useful if your team runs Claude in more than one surface, not just the terminal. It shows about 3.2k stars in the aitmpl.com directory. Like the other community hubs, it's a discovery layer: it organizes and points, it doesn't verify, scan, or benchmark anything, and it doesn't ship its own package manager. For a small business that has adopted Claude across the desktop app and the CLI, it's one of the few places that indexes both worlds in one location. For teams purely on Claude Code with a need for trust evidence, it adds breadth but not verification.

pricing See pricing

• + Indexes Claude beyond just Claude Code, including Desktop and the Agent SDK
• + One location for many extension types
• + Free and open on GitHub

• − Discovery only, with no verification or benchmarking
• − No install tooling of its own
• − Still Claude-centric, unlike cross-agent options

https://github.com/davepoon