Refactor CodeCanvas execution environment over sandbox primitives#136
Refactor CodeCanvas execution environment over sandbox primitives#136camerondurham wants to merge 8 commits intomainfrom
Conversation
![chatgpt-codex-connector[bot]](https://avatars.githubusercontent.com/in/1144995?s=60&v=4){kind=small}
There was a problem hiding this comment.
💡 Codex Review
Here are some automated review suggestions for this pull request.
Reviewed commit: cef204af86
ℹ️ About Codex in GitHub
Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you
- Open a pull request for review
- Mark a draft as ready
- Comment "@codex review".
If Codex has suggestions, it will comment; otherwise it will react with 👍.
Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".
engine/sandbox/linux/runner.go
Outdated
| sandbox.DebugPrintf("sandbox fallback to direct run after isolation error: %v", err) | ||
| return runDirect(ctx, input.Command, workDir) |
There was a problem hiding this comment.
When runIsolated fails (for example on hosts where unshare/user namespaces are not permitted), this path silently falls back to runDirect, which executes the user command as the server process. Because RuntimeAgent now uses linuxsandbox.NewRunner() by default, this non-strict fallback is active in normal execution and bypasses the previous process wrapper’s UID/GID and rlimit enforcement, so unsupported sandbox environments run code effectively unsandboxed instead of failing safely.
Useful? React with 👍 / 👎.
1 participant
No description provided.