Cordium: FOSS sandbox platform that eliminates credential injection
Octelium identity proxy eliminates credential injection in sandboxes.
Open source background agents
Credential proxying keeps keys out of sandboxes, unlike Ramp Inspect.
Engineering teams building internal AI agent workflows
Ramp Inspect · Stripe Minions · E2B
We saw larger tech companies like Ramp (Inspect) and Stripe (Minions) build this internally and thought an open source version should exist.
We made a few very intentional decisions when working on this:
1. Credentials are kept out of the sandbox. Authorized access goes through a proxy, so agents do not directly receive credentials.
2. The harness is not our problem. We're not going to tackle things like memory, self-learning.
3. No magic. Configurations are explicit. You can bring your own keys for models, sandboxes, and other providers. You can write your own instructions and agent.
Mistle can be run locally with a single command: https://github.com/mistlehq/mistle#run-mistle-locally
Questions, feedback and ideas are welcome!
Octelium identity proxy eliminates credential injection in sandboxes.
Thirty-line agent loop with Docker sandboxing contains blast radius safely.
MCP sandbox isolation for agents; E2B/Modal/Docker/WASM backends already exist separately.
Host-side credential enforcement for containerized agents before they leak your AWS keys.
Secretless infra access from sandboxes beats credential injection in Codespaces and E2B.
Agent credential proxy using OS keychain injection—credentials never touch agent memory.