Bounded shell and CLI execution for AI agents: structured contracts, policy-gated execution, hardened Linux runtime enforcement, and signed receipts.
-
Updated
Apr 13, 2026 - Python
Bounded shell and CLI execution for AI agents: structured contracts, policy-gated execution, hardened Linux runtime enforcement, and signed receipts.
Capability-security kernel for autonomous agents β seccomp/SELinux for agentic AI. Formal, auditable, language-agnostic, cryptographically verifiable.
A sample fastapi web app to integrate with a Tahoe-LAFS grid
Capability-based safety gates for LLM agents. The model proposes; signet authorizes.
Toy governance CLI demo: deny-by-default βdanger actionsβ gated by signed, expiring capability leases (global revoke-all + nonce revoke) plus guarded memory quarantine. Simulation-only: no real network/shell/files. Includes tripwire + tests to prevent misuse.
Autonomous loop agent that audits other agents for missing capability receipts
A simple example of core Tahoe-LAFS features using the REST interface.
Add a description, image, and links to the capability-security topic page so that developers can more easily learn about it.
To associate your repository with the capability-security topic, visit your repo's landing page and select "manage topics."