@viraatdas/rudder@2.14.81
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 9d ago · by lpm-firewall-aiThe scanner's similarity signals did not establish attack behavior in the inspected source. The flagged runtime locations serve local agent-data discovery or explicit Rudder Cloud workspace operations, while the install hook performs ordinary jj setup.
Decision evidence
public snapshot- The manifest runs scripts/postinstall.mjs after installation; that helper skips when jj is present and installs only the jj version-control tool when it is missing.
- Cloud launch and workspace attach stage a snapshot for an explicit Rudder Cloud operation and send it to the Rudder Cloud API; the default endpoint belongs to that service.
- The flagged cloud location stages Claude credentials locally; the handoff location identifies Claude's local projects directory, and the models location reads the local Codex model cache.
- Other flagged code describes an explicitly installed schedule, read-only session collection, local backend selection, a loopback-only board, update caching, feedback redaction, and a propose-first improvement default.
Source & flagged code
19 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@2.14.81.
See all versions of @viraatdas/rudder for other recorded verdicts.
- Review the evidence and your use of @viraatdas/rudder@2.14.81 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.