# Agentbeam Docs

> Practical guides for Beam Agent CLI, workspace policies, MCP tools, and browser protection.

- [Cookbooks](https://docs.agentbeam.com/cookbooks/overview.md): Step-by-step recipes that take you from nothing to using an AgentBeam feature fully, with commands to copy and a prompt to hand the whole recipe to your coding agent.
- [Agent monitoring, end to end](https://docs.agentbeam.com/cookbooks/agent-monitoring.md): Roll AgentBeam out to every developer machine: coding agents hooked and captured, MCP servers inventoried, a workspace policy moved from observe to enforce, alerts routed, and a weekly review routine.
- [App monitoring, end to end](https://docs.agentbeam.com/cookbooks/app-monitoring.md): Take a running app from unwatched to fully covered: traffic and logs watched for attacks, security events and LLM features reported, dependencies checked in CI, alerts routed, and a daily routine for what comes in.
- [Evaluate agents before you trust them](https://docs.agentbeam.com/cookbooks/agent-evaluation.md): Judge each agent action with a model or decision service you choose, test verdicts on sample actions, and benchmark a whole agent stack with WipeBench before giving it more autonomy.
- [Test skills, MCP servers and packages before agents use them](https://docs.agentbeam.com/cookbooks/skill-testing.md): Scan every skill before it's installed, gate risky ones at load time, keep an inventory with risk scores, fail CI on a dangerous skill, and check MCP configs and packages the same way.
- [Share skills and agent instructions with your team](https://docs.agentbeam.com/cookbooks/skill-sharing.md): Publish skills, CLAUDE.md and AGENTS.md, slash commands, subagents and Cursor rules to your workspace's team library, review them, and keep every teammate's agents on the same approved versions.
- [Next.js](https://docs.agentbeam.com/cookbooks/frameworks/nextjs.md): Monitor a Next.js app with the in-process SDK: requests in proxy.ts, sign-ins through Auth.js, LLM routes checked, and Next.js-specific attacks like the middleware bypass caught.
- [Express and Fastify](https://docs.agentbeam.com/cookbooks/frameworks/express.md): Monitor an Express or Fastify API with the in-process SDK, reporting each request after the response so every finding knows whether the attack worked.
- [NestJS](https://docs.agentbeam.com/cookbooks/frameworks/nestjs.md): Monitor a NestJS API with the in-process SDK, using a global middleware for requests and your auth service for sign-in events.
- [Django](https://docs.agentbeam.com/cookbooks/frameworks/django.md): Monitor a Django app: a request-logging middleware with status, auth signals for sign-ins, and beam monitor reading the logs with the Django rules on.
- [Flask](https://docs.agentbeam.com/cookbooks/frameworks/flask.md): Monitor a Flask app: an after-request hook logging status, sign-in lines from your login view, and beam monitor with the Flask rules, including the Werkzeug debugger console.
- [FastAPI](https://docs.agentbeam.com/cookbooks/frameworks/fastapi.md): Monitor a FastAPI service: an HTTP middleware writing one JSON line per request with status, sign-in lines from your auth dependency, and beam monitor reading them.
- [Ruby on Rails](https://docs.agentbeam.com/cookbooks/frameworks/rails.md): Monitor a Rails app: a Rack middleware writing one JSON line per request with status, Devise and Warden hooks for sign-ins, and beam monitor with the Rails rules.
- [Laravel](https://docs.agentbeam.com/cookbooks/frameworks/laravel.md): Monitor a Laravel app: a terminable middleware logging each request with status, auth event listeners for sign-ins, and beam monitor with the Laravel and PHP rules, Ignition included.
- [WordPress](https://docs.agentbeam.com/cookbooks/frameworks/wordpress.md): Monitor a WordPress site: beam monitor on the web server's access log with the WordPress rules, and a small must-use plugin that logs sign-ins and role changes.
- [Spring Boot](https://docs.agentbeam.com/cookbooks/frameworks/spring-boot.md): Monitor a Spring Boot service: Tomcat's access log with status, Spring Security events for sign-ins, and beam monitor with the Spring and Java rules, actuators included.
- [Go (net/http, chi, Gin)](https://docs.agentbeam.com/cookbooks/frameworks/go.md): Monitor a Go service: a net/http middleware logging one JSON line per request with slog, security events through slog, and beam monitor with the Go rules.
- [ASP.NET Core](https://docs.agentbeam.com/cookbooks/frameworks/aspnet-core.md): Monitor an ASP.NET Core app: a middleware logging one JSON line per request with Serilog, Identity sign-in events, and beam monitor with the .NET rules.
- [Agent CLI overview](https://docs.agentbeam.com/agent-cli/overview.md): Understand agent activity locally, scan the files your assistants use, and connect shared workspace controls when needed.
- [CLI quickstart](https://docs.agentbeam.com/agent-cli/quickstart.md): Install Beam, run beam setup, see your agents in the dashboard. Two commands, about two minutes.
- [Connect a workspace](https://docs.agentbeam.com/agent-cli/connect.md): Pair your machine through the browser or an enrollment code, then verify identity and policy synchronization.
- [Connect a Raspberry Pi](https://docs.agentbeam.com/agent-cli/raspberry-pi.md): Install Beam on a compatible Raspberry Pi, connect a workspace over SSH, and inspect agent activity from your laptop.
- [Connect coding assistants](https://docs.agentbeam.com/agent-cli/agent-hooks.md): Install native agent hooks, understand adapter coverage, and verify a real captured event.
- [Local Studio](https://docs.agentbeam.com/agent-cli/studio.md): Search captured activity, review findings, and scan content from the collector’s built-in browser interface.
- [Historical extraction](https://docs.agentbeam.com/agent-cli/extraction.md): Read existing Claude Code or Codex transcripts without a live hook, then optionally import the normalized records.
- [App monitoring and dependency checks](https://docs.agentbeam.com/agent-cli/app-monitoring.md): Watch a running application's logs and traffic around the clock, alert on attacks in Slack, PagerDuty and more, and check its dependencies for known vulnerabilities.
- [Scan skills & MCP configs](https://docs.agentbeam.com/agent-cli/scanning.md): Run offline heuristic checks on instruction files or MCP JSON, and save reports only when you choose.
- [Scan npm & PyPI packages](https://docs.agentbeam.com/agent-cli/package-scanning.md): Scan a package and its whole dependency tree for malware and known vulnerabilities before installing it, or audit what's already in node_modules and your virtualenvs.
- [Detection & custom rules](https://docs.agentbeam.com/agent-cli/rules.md): Inspect the active catalog, add focused local patterns, and understand session-level sequence findings.
- [Security skills & reviewers](https://docs.agentbeam.com/agent-cli/security-skills.md): Use Beam’s packaged review workflows for skills, MCP servers, AI assets, application boundaries, and incident evidence.
- [Optional Jev action judging](https://docs.agentbeam.com/agent-cli/jev.md): Opt into TypeSafe Jev checks with your own key, review Noul and Score results, and keep existing Beam workflows independent.
- [Judge actions with any LLM or decision service](https://docs.agentbeam.com/agent-cli/decision-models.md): Send redacted agent actions to a general model or your own decision service for an allow, review, or deny verdict, alongside or instead of Jev.
- [Dashboard & activity](https://docs.agentbeam.com/agent-cli/dashboard.md): Use the shared workspace to inspect connected devices, recorded agent activity, and the context behind findings.
- [Workspace policies](https://docs.agentbeam.com/agent-cli/policies.md): Set an organization baseline, add stricter member overrides, and verify what the installed CLI can enforce.
- [Policy presets](https://docs.agentbeam.com/agent-cli/policy-presets.md): Add ready-made restrictions for commands and tools, review the exact matches, and publish them to your workspace or account.
- [Risk review](https://docs.agentbeam.com/agent-cli/risks.md): Triage flagged activity, mark an assessment, and reopen a finding when new evidence changes the conclusion.
- [After an attack: responding to a finding](https://docs.agentbeam.com/agent-cli/incident-response.md): What to do when App monitoring or a device alert flags an attack: confirm it, contain it, clean up, and close it out, with a playbook for each kind of finding.
- [Integrations and alert workflows](https://docs.agentbeam.com/agent-cli/integrations.md): Connect Slack, Microsoft Teams, Discord, Google Chat, PagerDuty, a signed webhook or email, and build workflows that decide which alerts go where.
- [Members & roles](https://docs.agentbeam.com/agent-cli/members.md): Invite teammates, assign the smallest useful role, and manage workspace ownership and access.
- [Workspace & account settings](https://docs.agentbeam.com/agent-cli/workspace-settings.md): Manage invitation domains and future payload collection, and confirm your current account identity.
- [Audit log & CSV export](https://docs.agentbeam.com/agent-cli/audit.md): Inspect administrative actions and export a bounded, filtered record for your team’s review.
- [Subscription & enterprise](https://docs.agentbeam.com/agent-cli/subscription.md): Read seat usage and edition details, and separate public CLI capabilities from optional enterprise components.
- [Run in the background](https://docs.agentbeam.com/agent-cli/service.md): Keep the local collector running at login using macOS launchd or a Linux user service.
- [Configuration reference](https://docs.agentbeam.com/agent-cli/configuration.md): Configure local storage, collector pairing, and remote workspace destinations without mixing their roles.
- [Data, redaction & export](https://docs.agentbeam.com/agent-cli/data-privacy.md): Know what stays local, what an enrolled collector can send, and how to preserve evidence within retention limits.
- [Self-host the dashboard](https://docs.agentbeam.com/agent-cli/self-hosting.md): Plan a licensed independent installation and connect CLI devices to its own API and browser origin.
- [Beam MCP overview](https://docs.agentbeam.com/agent-cli/mcp/overview.md): Make Beam’s local scanning and inspection tools available to an MCP-compatible assistant.
- [Connect an MCP host](https://docs.agentbeam.com/agent-cli/mcp/setup.md): Build the local server, configure a stdio entrypoint, and verify tool discovery in your assistant.
- [MCP tool reference](https://docs.agentbeam.com/agent-cli/mcp/tools.md): Inputs, outputs, and operational limits for all five Beam MCP tools.
- [Browser extension](https://docs.agentbeam.com/agent-cli/extension/overview.md): Review risky prompts and commands before supported AI websites submit them.
- [Workspace browser data](https://docs.agentbeam.com/agent-cli/extension/workspace-data.md): Understand the workspace extension's browser events, metadata defaults, text collection controls, and rollout requirements.
- [Install the extension](https://docs.agentbeam.com/agent-cli/extension/install.md): Build and load Beam Sentinel from a reviewed checkout, then verify site access and local protection.
- [Review, redact & allow once](https://docs.agentbeam.com/agent-cli/extension/review.md): Handle a paused submission, inspect edited content, and understand what the local action history records.
- [CLI command reference](https://docs.agentbeam.com/agent-cli/commands.md): Find the exact command for setup, capture, scans, services, identity, rules, and removal.
- [Local collector API](https://docs.agentbeam.com/agent-cli/collector-api.md): Integrate with the authenticated loopback HTTP API for ingest, scans, review state, and redacted export.
- [Uninstall & offboarding](https://docs.agentbeam.com/agent-cli/uninstall.md): Revoke workspace access, preserve needed evidence, and remove the right local components.
- [CLI troubleshooting](https://docs.agentbeam.com/agent-cli/troubleshooting.md): Diagnose installation, local capture, pairing, forwarding, and policy behavior one stage at a time.