A token-efficient open-source AI assistant that remembers, adapts, and improves — secured, self-hosted, and entirely yours.
🛡️ Local credential safety and runtime armor for AI agents. Ephemeral container sandbox, sub-ms stream redactor, HTTP credential broker, canary tripwires, and MCP virtual proxy.
Multi-instance Docker daemon installer with sysbox-runc. Run isolated Docker instances side-by-side on the same host.
Container isolation for code you can't trust but have to run.
Self-hosted platform for running coding agents (Claude Code, Pi) in isolated containers. One task = one branch = one container; PRs are the only human approval gate.
Self-hosted database hosting and lifecycle management for PostgreSQL, MySQL, MariaDB, Redis, MongoDB, ClickHouse, and Qdrant—with backups, restores, monitoring, upgrades, and resource limits.
OS-enforced isolation demo for bounded local AI: an agent reaches a private vault only through the severino-vault-mcp sensitivity gate, with no filesystem bypass.
Isolated AI agent collaboration hub with HMAC auth and Docker sandbox enforcement
Ephemeral zero-credential sandbox for untrusted and agent-written code: isolates, resource limits, verified post-conditions, threat tests.
Run Claude Code agents in parallel, each in its own disposable Docker container with your repo mounted read-only. Reruns every builder's tests, flags out-of-scope edits and leaked tokens, merges onto one branch and verifies it with no network.
Policy-bound runtime execution and isolation engine for immutable AI agent skill artifacts
To associate your repository with the container-isolation topic, visit your repo's landing page and select "manage topics."