@viraatdas/rudder@3.0.3
A TUI for managing coding agents: run Claude Code, Codex and opencode side by side, each in its own isolated jj workspace.
AI Security Review
scanned 7d ago · by lpm-firewall-aiSource inspection found a postinstall hook limited to checking for and installing the package's required jj tool. Credential staging occurs in the explicit cloud-workspace flow, and the inspected agent-related files access expected local project or model data; no concrete attack behavior was identified.
Decision evidence
public snapshot- The postinstall hook checks for jj and, if absent, attempts its standard installation through Homebrew or Cargo.
- Claude credentials can be staged into a cloud snapshot, but the inspected path creates that snapshot during an explicit cloud operation and sends it to Rudder Cloud.
- The Claude project directory and Codex model cache are read from their expected locations under the user's home directory.
- Improvement collection reads registered project sessions into redacted records, while its scheduler is a separate command.
- Rudder state is stored under its own home directory; the board daemon uses a per-process token and loopback protections.
- The inspected update-check and feedback files describe local update state and feedback that excludes prompts, diffs, and file contents.
Source & flagged code
17 flagged · loading sourcePackage defines install-time lifecycle scripts.
package.jsonView on unpkgInstall-time lifecycle script is not statically allowlisted and needs review.
package.jsonView on unpkgSource file is highly similar to a previously finalized malicious package; route for source-aware review.
dist/util.jsView on unpkgFindings
Affected versions and remediation
This report applies to @viraatdas/rudder@3.0.3.
See all versions of @viraatdas/rudder for other recorded verdicts.
- Review the evidence and your use of @viraatdas/rudder@3.0.3 before allowing it. Restrict the permissions described in this report.
- Choose an independently verified alternative or release. This report does not establish that other versions are safe.
Evidence last updated: .
Think this verdict is wrong? Dispute it on GitHub or read how reports are produced.