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/SocketDev/sauceWrote 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/rules/socketdev/sauce/socket-mcp)<a href="https://agentmods.dev/rules/socketdev/sauce/socket-mcp"><img src="https://agentmods.dev/badge/rules/socketdev/sauce/socket-mcp/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/rules/socketdev/sauce/socket-mcp"><img src="https://agentmods.dev/badge/rules/socketdev/sauce/socket-mcp.svg" alt="Reviewed on agentmods" width="80" height="20"></a>What it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5.1 | $0.00035 | $0.00587 |
| Opus 5.5 | $0.00014 | $0.00235 |
| Sonnet 5.5 | $0.00007 | $0.00117 |
| Haiku 4.5 | $0.00003 | $0.00059 |
Grade A, and why
socket-mcp 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 yesterday.
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.
How it starts
The opening of the file, as written. The whole thing — 37 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Socket MCP
This plugin connects Cursor to Socket's hosted MCP server at https://mcp.socket.dev/. Cursor shows it as plugin-socket-skills-socket.
Sign in
The server uses OAuth. The user signs in to Socket in the browser. Nobody copies or pastes an API key.
When a Socket tool returns Unauthorized, or the server shows that it needs a login:
- Tell the user to open Customize in the Cursor sidebar, find the Socket MCP server, and choose Authenticate.
- When the user says the sign-in is done, try the tool one more time.
- If the tool still returns
Unauthorized, do not ask the user to sign in again. Some Cursor versions do not complete the sign-in for MCP servers that a plugin adds. Continue with the Socket CLI steps in the skills (pnpm dlx socket login), and tell the user that MCP checks are not available in this session.
Do not ask the user for an API key for this server. Do not put a token in an Authorization header for this server.
Tools
depscore: get Socket scores (0-100) for one or more packages. Give each package asdepname,ecosystem(defaultnpm), andversion(use the exact version when you know it). Lower scores mean higher risk. A missing score is not a safety verdict.organizations: list the Socket organizations that the sign-in can access. Use it to find theorg_slugthatalertsandthreat_feedneed. If there is more than one organization, ask the user which one to use.alerts: list the security alerts of an organization. It needsorg_slug. Filter withstatus,severity, and other fields, and usecursorto get more pages.threat_feed: list packages that Socket flagged recently, for example malware and typosquats. It needsorg_slug.package_files,package_file_contents,package_file_grep: read the files in a published package before you install it.
When to check
- Before you add a new dependency, run
depscoreon it. Report low scores to the user before you continue. - Before you upgrade a dependency, run
depscoreon the target version. - When the user asks if a package is safe, use
depscore, then the/socket-inspectskill for a full review.
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.
- yesterday First seen · 37 lines · 35 tokens per session scan A 2c1dfc1f2be1
socket-mcp is a cursor rule published in the GitHub repository SocketDev/sauce (3 stars, last pushed 2d ago), licensed MIT. It adds 35 tokens to every session and 587 once invoked, about $0.0001 per session on Opus 5.5. 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 2026-11-07.
socket-mcp compared with cursor rules from other repositories
| Mod | Type | Stars | Last push | Licence | Scan | Tokens | Works with |
|---|---|---|---|---|---|---|---|
| socket-mcpSocketDev/sauce | Cursor rule | 3 | 2d ago | MIT | A | 35 | Cursor |
| ponytailDietrichGebert/ponytail | Cursor rule | 186k | 10d ago | MIT | A | 604 | Cursor |
| angular-20angular/angular | Cursor rule | 101k | yesterday | MIT | A | 0 | Cursor |
| dev-standardapache/superset | Cursor rule | 75k | yesterday | Apache-2.0 | A | 1,147 | Cursor |
| control-plane-descriptorsprisma/orm | Cursor rule | 48k | yesterday | Apache-2.0 | A | 0 | Cursor |
| family-instance-domain-actionsprisma/orm | Cursor rule | 48k | yesterday | Apache-2.0 | A | 0 | Cursor |
| prefer-assertions-over-defensive-checksprisma/orm | Cursor rule | 48k | yesterday | Apache-2.0 | A | 0 | Cursor |