Secure everything
your AI touches.
AI agent security for coding agents, AI chats and the LLM apps you ship.
npx @agent-beam/beam add skill agentbeam-setup -g --all-agentsNo credit card required · Open-source CLI · Cloud or self-hosted
Findings
Last 24 hours · every surface
- Open findings · 3 critical
- 12
- Blocked today · enforce policy
- 48
- Secrets redacted · before send
- 17
- Apps monitored · watch-only
- 7
- Prompt injection in hidden Unicodeapi-server · POST /v1/chatCritical
- Read ~/.aws/credentialsClaude Code · billing-svcFlagged
- Live AWS key removed before sendChatGPT · SentinelRedacted
- rm -rf ./ deniedCodex · infraBlocked
- Package on the CISA KEV listbeam cve check · CIBuild failed
- Hidden instruction in tool descriptionMCP · before connectHeld
Works with the AI tools your teams already run
- Claude Code
- Codex
- Cursor
- GitHub Copilot
- ChatGPT
- Gemini
- Node.js
- Next.js
One AI security platform
Every surface AI touches, in one record.
- AppsNew
Prompt injection and attack detection for the apps you ship
Watches logs, traffic and LLM features. Never blocks a request, and raw logs never leave your server.
- Injection
- Prompt injection
- SSRF
- Reverse shells
- Path traversal
- Leaked keys
- Agents
AI coding agent monitoring
Every command, file and tool call, with policy that can say no.
- Browser
Data leak prevention for AI chats
Secrets and sensitive pastes caught before they reach ChatGPT, Claude or Gemini.
- Supply chain
MCP, skill and package scanning
Skills, MCP servers and npm or PyPI packages scanned before they install. Lockfiles checked in CI.
- Team libraryNew
Share skills, CLAUDE.md and subagents
Publish once, install on every machine. Each share is scanned, versioned and, if you want, reviewed.
- Device managementApple and Windows devices, enrolled and configured from the same workspace.Coming soon
How it works
AI agent security, live in minutes.
Install with one command
Or ask your coding agent to do it. The setup skill is scanned before it installs.
Watch every action
Agents, AI chats and apps recorded as they work. Secrets are redacted on the device.
Enforce and prove
Observe, advisory or enforce policy, with alerts to Slack and a hashed audit export.
$ npx @agent-beam/beam add skill agentbeam-setup -g --all-agents
✔ scanned agentbeam-setup · no findings
✔ installed for Claude Code, Codex, Cursor
› set up AgentBeam
✔ hooks installed for 3 agents
✔ collector listening on 127.0.0.1:4319
✔ connected to workspace acme · policy: advisory
watching. findings appear at app.agentbeam.com
Enterprise ready
Built to pass your security review.
Least-privilege roles
Owner, administrator, security manager, auditor, billing and member.
Administrative audit trail
Every invite, role and policy change recorded, exportable as CSV.
Policy you set
Observe, advisory or enforce, chosen by owners and admins.
Redacted at the source
Secrets removed on the device before anything leaves it.
Self-host the control plane
Docker, your PostgreSQL and an offline, installation-bound license.
Open-source core
Read the CLI that runs on your machines before you trust it.
AI threat intelligence
Agentic AI attacks and CVEs, tracked so you don’t have to.
- CVEs tracked
- 5,477
- On CISA KEV
- 1,756
- New in 30 days
- 256
Latest
- Oct 23Agent incidentPanasonic HR data sent to generative AI
- Oct 23Agent incidentMythos 5 malicious pull request (AISI test)
- Oct 23Agent incidentOpenAI containment red team
- Oct 22AI-run attackSymantec "Skynet mode"
- Oct 25ResearchBackdoored Open-Weight Qwen Model Steals .env Files Through Codex
- Oct 25ResearchMalicious AGENTS.md in Downloads: How a Planted File Hijacks Codex and Claude Code
Pricing
Free for one developer. Ready for the whole org.
Individual
One developer
Free
- Agent activity monitoring
- Skill, MCP and package scanning
- Sentinel Chrome extension
- Local-first, no account
Team
Teams and companies
$10 /seat/mo
- Six workspace roles
- Observe, advisory or enforce policy
- App monitoring and CVE checks
- Team library for skills and CLAUDE.md
- Audit trail with CSV export
On-premise
Regulated environments
Custom
- Self-hosted control plane
- Your PostgreSQL, Docker deploy
- Offline, installation-bound license
- Seat allocation and access controls
What is Agentbeam?
Agentbeam is an AI agent security platform. It records what AI coding agents do on developer machines, checks what people send to AI chats in the browser, watches the applications you ship for attacks, and scans skills, MCP servers and packages before they run. Findings arrive in one workspace with the evidence attached and an audit trail.
What does App monitoring do?
App monitoring watches a running application's logs, HTTP traffic and LLM features for attacks such as injection, path traversal, SSRF, reverse shells, prompt injection and attempts to extract API keys. It runs as the open-source beam CLI beside your app or as an SDK inside a Node.js or Next.js app. It only watches: it never blocks requests, and raw log lines stay on your server.
Does Agentbeam send my data anywhere?
Not by default. On developer machines, activity is recorded locally and nothing leaves unless you connect a workspace or configure an export. App monitoring sends only redacted findings, never raw log lines, and the Sentinel Chrome extension makes no network request at all.
Does Agentbeam block risky agent actions?
It can. Workspace owners and admins set observe, advisory, or enforce policy. Enforce denies matching tool calls on Claude Code and Codex. The Sentinel Chrome extension pauses or blocks high-risk browser sends and can hold a send after redacting live secrets. Cursor, Copilot CLI and Gemini currently get a deny note, not a verified block. App monitoring never blocks.
Can Agentbeam scan skills, MCP servers and packages?
Yes. beam add skill scans a skill before installing it and beam skills scan checks the ones already installed. beam scan --mcp checks an MCP config for tool poisoning, unpinned versions that enable rug pulls and credential-exfiltration phrasing. beam add package scans an npm or PyPI package and its whole dependency tree for install-time scripts, environment theft and typosquats, and installs only if nothing critical turns up; beam packages scan audits what's already installed. beam cve check matches lockfiles against Beam's CVE database and can fail a CI build. Scanning is heuristic triage, not a guarantee of safety.
Can my team share skills and CLAUDE.md files?
Yes. beam share publishes a skill, a CLAUDE.md, AGENTS.md or GEMINI.md, a slash command, a subagent or a Cursor rule to your workspace's Team library, and teammates install it with beam share install. Every share is scanned when it's published and again before it installs, versions are kept, and admins can require review of risky shares or every share.
Which AI agents and tools does Agentbeam work with?
Coding agents including Claude Code, Codex, Cursor, GitHub Copilot, Antigravity and Muse; AI chats such as ChatGPT, Claude and Gemini through the Sentinel Chrome extension; and Node.js and Next.js apps, or any app that writes logs, through App monitoring.
Who is Agentbeam built for?
Individual developers who want to see what their agents do, free and local-first. Security teams that need detection and evidence across every agent, browser and app in use. IT teams rolling out AI tools who need workspace roles, policy and an audit trail, in the cloud or self-hosted.
How do I get started with Agentbeam?
Click Start free to sign in and set up your workspace, or run npx @agent-beam/beam add skill agentbeam-setup -g --all-agents and ask your coding agent to set up AgentBeam. Rolling out to a team? Book a demo, or email [email protected].