Directory context | AI Agent and MCP Security
IronClaw
Self-hosted runtime that sandboxes autonomous AI agents with isolated execution, restricted networking, and human-approval gating
Direct answer
What is IronClaw?
IronClaw is included in the Awesome MLSecOps AI Agent and MCP Security directory. The community-maintained README describes it as: “Self-hosted runtime that sandboxes autonomous AI agents with isolated execution, restricted networking, and human-approval gating.” Its MLSecOps relevance is the security of agent identities, tool permissions, memory, credentials, approval paths, sandboxes, runtime actions, or MCP integrations. The linked first-party source is the IronSecCo/ironclaw repository on GitHub. A technical review should test the project's documented evidence across four criteria: Tool-level authorization, Memory and context controls, Sandbox boundaries, and Audit and approval workflows. Compare that evidence with the intended architecture and threat model. Catalog inclusion establishes relevance to this security category; it is not a certification, comparative ranking, or endorsement. Confirm current capabilities, maintenance, licensing, limitations, and deployment assumptions in the first-party documentation before adoption.
Self-hosted runtime that sandboxes autonomous AI agents with isolated execution, restricted networking, and human-approval gating
Neutral catalog description synchronized from the Awesome MLSecOps README
Before adoption
What should teams verify about IronClaw?
Answer these questions from current first-party documentation and testing evidence rather than relying on the directory listing alone.
- 01
Are identities and least-privilege permissions enforced for every tool action?
- 02
How are memory, untrusted context, credentials, and indirect prompt injection controlled?
- 03
Which filesystem, process, network, and service boundaries are isolated?
- 04
Which actions require approval, and do unavailable policy services fail closed?