Context7 is a documentation service that puts current, version-specific library documentation and code examples into prompts for large language models and AI coding editors. Developers use it to help coding agents produce answers based on the libraries they actually use; the catalogue entries provide ways for agents to access Context7 through its CLI or MCP server.
Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
git clone --depth 1 https://github.com/upstash/context7Wrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/plugins/upstash/context7/gemini-extension)<a href="https://agentmods.dev/plugins/upstash/context7/gemini-extension"><img src="https://agentmods.dev/badge/plugins/upstash/context7/gemini-extension/github.svg" alt="Measured on agentmods" height="20"></a>Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.
<a href="https://agentmods.dev/plugins/upstash/context7/gemini-extension"><img src="https://agentmods.dev/badge/plugins/upstash/context7/gemini-extension.svg" alt="Reviewed on agentmods" width="80" height="20"></a>Grade A, and why
context7 scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 1mo ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
What it actually says
{
"name": "context7",
"description": "Up-to-date code docs for any prompt",
"version": "1.0.0",
"settings": [
{
"name": "API Key",
"description": "Context7 API key. Create new key at https://context7.com/dashboard.",
"envVar": "CONTEXT7_API_KEY"
}
],
"mcpServers": {
"context7": {
"command": "npx",
"args": ["-y", "@upstash/context7-mcp", "--api-key", "${CONTEXT7_API_KEY}"]
}
}
}
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- 1mo ago First seen · 19 lines scan A cfc93e200311
context7 is a plugin published in the GitHub repository upstash/context7 (63,103 stars, last pushed today), licensed MIT. Its token cost is not measured: this kind of file is read by the harness, not the model. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-09-08.
Other plugins, from other repositories
hyper-mcp
A fast, secure MCP server that extends its capabilities through WebAssembly plugins.
livetennisapi
Real-time and historical tennis data for Claude — live scores, match state (server, break-point, retirement/walkover), players, rankings, fixtures, head-to-head and the 1968-2022 results archive across ATP, WTA, Challenger, ITF and juniors. Connects to the Live Tennis API (free tier available, no card); collects no…
mcp-zap-connect
Guided Docker Compose setup for local MCP ZAP Server in Codex or Cursor.
umt
28 production-ready MCP servers for coding agents — filesystem, GitHub, npm, research, and more. Ships zero-config servers out of the box and the umt CLI to connect any harness.
declaude
Rewrites Claude-English into plain English: sycophantic openers, hedging filler and padded phrasing out, facts, code blocks, names and numbers untouched. Runs an open-weights model on dedicated GPUs, so text never reaches a commercial AI provider.
cortex
Cortex observability and investigation skills for an existing service.