beam-mcp
This schema exposes only Beam's local, offline, read-only scanning tools — nothing that changes state.
Scan inline text (
beam_scan_content): run Beam's heuristic risk scanner over raw content (up to 2M chars) with no network call and no collector running; returns findings with severity, title, explanation and matched evidence.Scan a file on disk (
beam_scan_path): same scan against a local file path, strictly read-only — the file is never modified.Pick a target
kind:skill(agent instruction file, default) ormcp(MCP server config JSON, which also checks dependency version pinning).Inspect detection rules (
beam_list_rules): see every active rule by category — built-in, custom from~/.beam/rules.json, and cross-event sequence rules — so you can judge what a scan result actually means.List hookable agents (
beam_list_supported_agents): which coding agents Beam can capture (claude-code, codex, cursor…), whether each hook config is installable, and whether its payload shape is verified.Check collector health (
beam_collector_status): whether the local collector is running on 127.0.0.1, plus mode, event count and retention; it never starts, stops or configures anything.
Note: everything here is read-only and local. The write tools the README describes (policy changes, risk marking, approval resolution) are not part of this schema.
Click on "Deploy Server".
Wait a few minutes for the server to deploy. Once ready, it will show a "Started" state.
In the chat, type
@followed by the MCP server name and your instructions, e.g., "@beam-mcp Scan this file for risky AI agent patterns"
That's it! The server will respond to your query, and you can continue using it as needed.
Here is a step-by-step guide with screenshots.
beam-mcp
An MCP server that exposes Beam's local, offline risk-scanning to any MCP-compatible host (Claude Desktop, Claude Code, and others).
Beam observes AI agent activity and heuristically flags risky patterns (secrets, destructive commands, exfiltration, unpinned MCP dependencies, etc.) — it never executes, blocks, or approves anything. This server is a thin, read-only wrapper around the beam CLI: it doesn't reimplement Beam's detection logic, and it doesn't add any capability the CLI doesn't already have.
Prerequisites
Beam CLI installed and on
PATH(npm install -g beam).Node.js >= 20.17.
Related MCP server: meok-mcp-injection-scan-mcp
Install & run
npm install
npm run buildPoint your MCP host at the built binary over stdio (the only transport this server supports today):
{
"mcpServers": {
"beam": {
"command": "node",
"args": ["/absolute/path/to/beam-mcp/dist/index.js"]
}
}
}Or, once published: npx beam-mcp.
Status
We're still building out the MCP layer for Beam. This first slice is intentionally narrow — local, read-only tools. A hosted version will be available at mcp.agentbeam.com.
Tools
All tools are read-only. None of them start, stop, configure, or send data anywhere beyond your own machine (and your own already-running Beam collector, for the one tool that talks to it).
Read tools (any connection)
Tool | What it does |
| Signed-in user, workspace, role and permissions. |
| "What needs attention": open and new risks by severity, top rules, agents and people, most flagged sessions, pending approvals, policy mode. Window 24h, 7d or 30d. |
| Rules reviewers keep dismissing, with dismiss rate and example findings: the likeliest false positives. |
| A clickable link to any dashboard page, for things without a tool (members, devices, usage, billing...). |
| Link only. Audit logs are not readable through assistants; this returns the dashboard link. |
| Effective policy for you, a member, or the workspace default. |
| Recent workspace policy versions and who published them. |
| Risk findings, and one finding explained in plain English. |
| Agent sessions, and the facts needed to summarize one. |
| Pending or resolved agent permission requests. |
Write tools (only on a Read-and-write connection)
Tool | What it does |
| observe / advisory / enforce. |
| Block a tool or disable an agent, workspace-wide or for one person. |
| Add, replace or remove a custom rule. |
| Republish an earlier policy version as a new one. |
| Accept a risk, mark it a false positive (optionally dismissing repeats), block it with a policy, suspend the agent, or mark it reviewed or reopen it. |
| Approve (10-60 min) or deny an agent request. |
Every write is a preview first: the tool changes nothing until it is called again with confirm: true. A write
goes through only if all of these hold, each checked by the dashboard on every call:
the connection was approved with Read and write (token scope
mcp:write);the person is a workspace Owner or Admin, and still has the underlying permission;
an Owner has switched on AI assistant changes in Dashboard → Workspace (off by default, and the emergency off switch);
it is within the write rate limit.
Applied changes are recorded in the audit log as made through an AI assistant, naming the app.
Read results include a url for each item so the person can click through to it in the dashboard.
The local scanning tools above are stdio-only and are not exposed remotely.
Security
Security is a core design constraint of this server, not an afterthought — it runs locally, exposes read-only tools only, and takes a conservative approach to anything touching your machine or the Beam collector. See SECURITY.md to report a concern.
Development
npm run dev # run src/index.ts directly with tsx
npm run typecheckLicense
AGPL-3.0-only — see LICENSE. Same license as beam-cli, for the same reason: this stays free and open, including for anyone who modifies it and offers it as a hosted service.
Related
beam-cli — the CLI and SDK this server wraps. Open source.
Available Tools
5 toolsbeam_collector_statusCheck the local Beam collector's statusA
Check whether Beam's local collector is running (it only ever binds to 127.0.0.1) and return its health summary: mode, event count, retention limit. Does not start, stop, or configure anything. Uses the same local credentials as the beam CLI itself — this tool never generates, stores, or exposes them.
Inputs: no named parameters.
beam_list_rulesList Beam's detection rulesA
List every active Beam detection rule, grouped by category: built-in rules, any custom rules loaded from ~/.beam/rules.json, and cross-event sequence rules. Use this to know what patterns Beam actually looks for before relying on a scan result. Read-only, runs entirely locally.
Inputs: no named parameters.
beam_list_supported_agentsList agents Beam can hook intoA
List the coding agents Beam knows how to capture activity from (e.g. claude-code, codex, cursor), whether each one's hook config is installable, and whether its hook payload shape has been verified against that agent's own docs. Read-only, runs entirely locally.
Inputs: no named parameters.
beam_scan_contentScan text for risky agent instructionsA
Run Beam's offline heuristic risk scanner over inline text — an agent skill/instruction file or an MCP server config (JSON). Runs entirely locally: nothing is sent over the network, and the collector does not need to be running. Returns a list of findings (severity, title, explanation, matched evidence).
Inputs: content (required), kind.
beam_scan_pathScan a local file for risky agent instructionsA
Run Beam's offline heuristic risk scanner over a file already on disk — an agent skill/instruction file or an MCP server config. Read-only: the file is never modified. Runs entirely locally.
Inputs: path (required), kind.
Tool Schema Changelog
Recent tool additions, removals, and schema changes observed during successful MCP inspections.
Loading tool schema changelog…
TDQS
Scored across 5 tools
Each tool targets a clearly distinct operation: scanning inline content, scanning a file, listing rules, listing supported agents, and checking collector status. There is no meaningful overlap or ambiguity between them.
All tools share a consistent beam_ prefix and most follow a verb_noun pattern (scan_content, scan_path, list_rules, list_supported_agents). beam_collector_status is a minor deviation since it uses a noun phrase rather than an explicit verb, but it remains predictable.
Five tools is well-scoped for a focused read-only security scanning utility. Each tool earns its place and there is no bloat or redundancy.
The tool surface covers the core read-only workflow: scanning content or paths, understanding available detection rules, knowing which agents are supported, and checking collector health. Since the server is explicitly read-only, there are no obvious missing operations.
Maintenance
Related MCP Connectors
Scan AI agent skills, MCP server packages and MCP configs before you install them. Static analysis.
Security scanner for MCP servers. Detect vulnerabilities, prompt injection, and tool poisoning.
Security scanner for AI agents and MCP servers: injection, jailbreak, tool abuse.
Scans an MCP server for security gaps before your agent connects. Free tier, no signup.
Related MCP Servers
- AlicenseNot gradedqualityBmaintenanceScans MCP servers for prompt injection, supply chain attacks, excessive permissions, and code execution risks. Includes an offline blacklist that catches known-compromised packages like LiteLLM 1.82.7/1.82.8 and Trivy with zero latency.19MIT
- AlicenseAqualityDmaintenanceScans MCP servers for prompt-injection, tool-poisoning, and SSRF vulnerabilities using 30+ canonical rules across 5 severity tiers, with optional signed safety reports for procurement.5MIT
- AlicenseCqualityDmaintenanceSecurity scanner and MCP server that catches dangerous patterns in MCP servers and AI agent projects, such as leaked secrets, shell execution, and prompt-injection text. Runs as both a CLI and MCP server with CI-friendly severity gates.21MIT
- AlicenseNot gradedqualityDmaintenanceScans MCP servers for security vulnerabilities, prompt injection, and tool poisoning, providing risk scores and protection.4MIT