AI Agents That Work With Claude (2026)
As of Aug 10, 2026, 216 AI agents in the KanonAgent index are judged to connect to Claude; this page covers the top 14 by real traction (14 with the source line quoted). Judgements come from quoted evidence, not vendor claims — where we cannot read it, we leave it blank.
Connecting an agent to Claude via Claude Code turns the model into a controllable coding and research environment with terminal access, plugins, and multi-agent loops. The 2026 shift is from single-prompt chats to persistent, skill-augmented sessions that run locally or across teams. Only agents with explicit Claude Code or MCP ties are included.
Updated 2026-08-10 · 14 products · live data from KanonAgent
1. ai-berkshire15.3k upvotes
ai-berkshire runs multi-agent value-investing analysis using four master methodologies, and the Claude Code integration lets you execute parallel adversarial research inside the Claude session.
source: “built for Claude Code / Codex”
2. reverse-skill14.7k upvotes
reverse-skill auto-selects and configures security toolchains for vulnerability research, and the Claude Code support lets you launch reusable exploit workflows directly from the Claude prompt.
source: “Supports Claude Code, K”
3. pm-skills10.0k upvotes
pm-skills supplies 100+ agentic plugins for product discovery and execution, and the Claude Code design lets you assemble full-lifecycle PM workflows without leaving the editor.
source: “Designed for Claude Code and Cowork”
4. ralph-claude-code9.6k upvotes
ralph-claude-code adds autonomous exit detection to Claude Code loops, stopping infinite code generation while letting the agent complete logic blocks on its own.
source: “Ralph for Claude Code”
5. claude-video8.4k upvotes
claude-video downloads, transcribes and summarizes any video for Claude, and the Claude Code integration feeds the extracted content straight into analysis or Q&A tasks.
source: “Claude Code (recommended”
6. humanizer6.2k upvotes
humanizer rewrites AI text to pass human detectors, and the Claude Code plugin lets you apply it automatically to any generated output before export.
source: “Claude Code users can also install Humanizer as a plugin”
7. openwork5.2k upvotes
openwork provides an open-source team workflow layer for Claude Code, enabling multi-role task orchestration across Codex, Cursor and other MCP clients.
source: “Add one OpenWork MCP to Codex, Claude Code, Cursor”
8. claude-seo3.0k upvotes
claude-seo bundles 25 SEO sub-skills and 18 sub-agents, and the Claude Code plugin runs technical audits and content optimization end-to-end inside the session.
source: “SEO analysis plugin for [Claude Code]”
9. DesktopCommanderMCP2.9k upvotes
DesktopCommanderMCP grants Claude terminal, file search and diff-editing rights, turning Claude Desktop into a full local development controller.
source: “with Claude Desktop and other MCP clients”
10. agents-cli2.3k upvotes
agents-cli deploys and evaluates agents on Google Cloud, and its Claude Code compatibility lets you build, test and ship from the same Claude prompt.
source: “Works seamlessly with: Antigravity CLI • Claude Code • Codex”
11. designer-skills2.0k upvotes
designer-skills supplies research-to-delivery design plugins, and the Claude Code integration lets the agent execute complete UI and systems design cycles.
source: “for Claude Code and Gemini CLI”
12. stitch-skills1.6k upvotes
stitch-skills offers reusable agent skills across coding platforms, and the Claude Code compatibility lets you share standardized commands between Claude, Codex and Gemini.
source: “Codex”
13. NanoResearch1.5k upvotes
NanoResearch autonomously handles literature review, experiment design and reporting, and the Claude Code mode lets you run the full research loop inside Claude.
source: “Claude Code 模式”
14. agmsg1.4k upvotes
agmsg lets Claude Code, Codex, Gemini and Copilot exchange code and results in one CLI, removing manual copy-paste between agents.
source: “Claude Code, Codex, Gemini & Copilot”
Before you wire it up
Check whether the agent ships as a native Claude Code plugin, MCP server, or CLI wrapper; deeper integrations expose terminal or file-system scopes while shallow ones only add prompt templates. Verify the exact permission model—DesktopCommanderMCP and similar tools request broad local access that may exceed what you want to grant. Rate limits appear when multiple sub-agents run in parallel; test with small tasks first. The most common breakage is version drift between Claude Code updates and plugin manifests, so pin the Claude version before installing skills.
FAQ
How do I install a Claude Code skill from this list?
Most agents provide a one-line MCP or plugin install command that you paste into Claude Code settings; restart the session to load the new tools.
What happens if an agent needs terminal access?
Agents such as DesktopCommanderMCP request explicit MCP permissions; you approve or deny each scope before the first run.
Can I mix skills from different agents in one Claude session?
Yes—stitch-skills and openwork are designed for cross-agent reuse, so you can load multiple compatible plugins together.
How is this list ranked?
By real traction from our index (upvotes / MRR / growth rate, whichever the product actually has) — not editorial picks, and we do not accept paid placement. Sources: ProductHunt, Hacker News, GitHub, HuggingFace, Reddit, TrustMRR.
What is the inclusion bar?
A page is published only when at least 5 real products qualify. Judgement fields (autonomy, prerequisites, cost, integrations) all require a source quote — where we cannot read it, we leave it blank rather than guess.
How often is this updated?
Collection runs continuously; this page was regenerated on 2026-08-10. Every number is verifiable through our public read-only API: https://kanonagent.com/data