New: the Agent Incident Ledger: disclosed incidents of AI gone rogueExplore the ledger

Discover every agent on endpoints, in browsers and in the cloud.

Get a full inventory of every agent, including the shadow AI living in your environment. Understand the full supply chain of MCP servers, skills, plugins and extensions, the moment they appear.

Manifold's inventory: agents and supply-chain components by risk, with the people and endpoints running them.

Every agent your people use, in one inventory.

100+ agent types, from coding agents like Claude Code, Codex and Cursor to desktop and browser assistants. Each one linked to the identity of its user, including when signed in with personal accounts.

Browser use: covered

Every session of agents used in the browser, or when the browser is driven by an agent.

See the agents running in the cloud.

Uncover the agents running in the cloud, such as Claude Code, Codex and OpenClaw.

Control what every agent runs on

Understand how each agent is set up, and what it depends on.

See when an agent runs with its safety controls switched off. The asset graph maps every skill, sub-agent and tool it relies on, and Manifold stops other agents from modifying it.

Vet every skill, plugin and MCP server.

Manifest, our research and threat intelligence on the agent supply chain, checks every component your agents load and re-checks every new version, across more than 500,000 skills, MCP servers, extensions and plugins. If one turns malicious, remove it from one endpoint or from all endpoints at once.

Developers keep the tools they like, and security knows which ones to trust.

AssessUncover risky configurations and actions, and cut exposure.

See Manifold in action.

A short tour of the platform, from the first agent discovered to the first policy enforced.