Back to browse
GitHub Repository

Open-source, general-purpose sandbox platform for devs and AI agents that provides identity-based secure access to infrastructure without credentials.

38 starsTypeScript

Cordium – FOSS identity-based sandbox platform with zero-trust access

by geoctl·Jun 14, 2026·2 points·0 comments

AI Analysis

●●●BangerBold BetSolve My ProblemNiche Gem

Secretless infra access from sandboxes beats credential injection in Codespaces and E2B.

Strengths
  • Identity-proxy architecture eliminates credential injection into sandboxed environments entirely.
  • Unified platform handles both human dev sessions and AI agent workloads with same security model.
  • Kubernetes-native with rootless containers, persistent or ephemeral workspaces, multiple access methods.
Weaknesses
  • Requires Octelium dependency, adding another system to deploy and maintain alongside Kubernetes.
  • Self-hosted Kubernetes barrier excludes smaller teams who'd prefer managed or simpler setups.
Target Audience

DevOps engineers, security teams, and developers managing sandboxed environments with infrastructure access needs

Similar To

GitHub Codespaces · E2B · Daytona

Post Description

Cordium is a FOSS, self-hosted, identity-based, general-purpose sandbox platform that I've been working on for a long time now that is built on Kubernetes and Octelium, my main project. The key differentiator here for Cordium, when compared to other dev environments (e.g. GitHub Codespaces) and sandbox platforms (e.g. E2B, Daytona, etc.), is that Cordium automatically provides identity-based, secretless secure access to resources/infrastructure (e.g. APIs, SSH, databases, k8s, etc.) without having to inject credentials (e.g. API keys, SSH private keys, database passwords, etc.) into the sandbox where the upstream credential is held by the identity-aware proxy of the Octelium-protected resource outside the reach of the sandbox.

In short, Cordium is not just an isolated execution environment that can replace remote development environments and sandbox platforms, but also equally a secure access platform to infrastructure/resources. It's basically a sandbox platform + a ZTNA/remote-access-VPN baked-in with unified identity management, L7-aware access control and visibility.

Similar Projects

Agntor – Trust infrastructure for AI agents (identity, escrow, guard)

The SDK exposes the exact primitives you want for autonomous-agent commerce: register/resolve identity, attest and badge verification, create/fund/cancel escrows, release/slash settlements, and reputation queries — plus event hooks. It's a smart, timely idea to stitch payments and trust into agent URIs, but the repo still reads like an early SDK: docs and integration examples are thin and there's no clear public security/settlement audit or adoption evidence yet.

Bold BetNiche Gem
zeshama
214mo ago