Menu
Blog How it works Use Cases
agentsh
Open-source runtime for AI agent security
Beacon
AI endpoint visibility and control
Watchtower
Enterprise control plane for agentsh and Beacon
Request Access
Unsupervised Agent Runtime Security

Runtime security for
headless agents.

Headless agents in CI, containers, and dev environments execute fast. AgentSH intercepts system calls and evaluates them against policy locally before they run. Allow, block, prompt, or redirect. You decide.

  • Execution-time enforcement Decisions happen at the moment of action, not after.
  • Built for unsupervised workflows No UI required. No user in the loop needed.
  • Full context awareness See the complete chain from agent to syscall.
  • Open source Single binary. Runs anywhere.
agentsh
$ agentsh -- agent-runner
Wrapping agent-runner with policy...
connect api.github.com:443 ALLOW
exec terraform apply PROMPT
connect pastebin.com:443 BLOCK
connect registry.npmjs.org REDIRECT
Four actions

More than just allow and block.

Allow. Block. Prompt. Redirect. Steer agents toward approved alternatives instead of forcing retry loops.

Allow

Action proceeds immediately. No user interaction needed.

Block

Action is denied. Agent receives an error response.

?

Prompt

User must approve before the action can proceed.

Redirect

Action is transparently routed to an approved alternative.

Why agentsh

Built for unsupervised agentic workloads.

Syscall-level interception

AgentSH intercepts network connections, file operations, and process execution before they happen.

Process chain tracking

Write policies that understand context, not just actions.

Policy as code

Define policies in YAML or JSON. Version in git. Test in CI. Deploy with confidence.

Single binary, no dependencies

No background services required. No cloud connection required. Works in air-gapped environments.

Use cases

Common places teams deploy AgentSH.

CI and build agents

Agents that fetch dependencies and run commands under policy, so builds stay safe without slowing down.

Codegen and refactor bots

Agents that touch repos and credentials, constrained to approved actions and destinations.

Ops agents

Agents that run infra actions (terraform, kubectl) under policy so blast radius is bounded.

Sandbox runners

Environments where blast radius must be bounded by default, with no human in the loop.

Platform

Use standalone or scale with the platform.

AgentSH works great on its own. When you need centralized policy, approvals routing, audit logs, and a kill switch, add Watchtower. If your risk is supervised AI on endpoints, use Beacon.

Beacon

Protects supervised AI on employee endpoints — macOS and Windows. Visibility, policy, and human-in-the-loop approvals for desktop AI tools.

Learn about Beacon →

Watchtower

Centralized policy, approvals routing, audit logs, and a kill switch across your entire AI fleet — both supervised and unsupervised.

Learn about Watchtower →
See use case: Headless agents in CI and containers Learn about Execution-Layer Security

Start securing your agents today.

AgentSH is open source and free to use. Clone the repo or read the docs.