@exla/rudder@3.9.5
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 20h ago · by lpm-firewall-aiThe automatic first-party agent skill setup is a concrete residual configuration risk, but the inspected source does not establish an attack or unrelated credential transfer. The unresolved runtime operations resolve to ordinary data handling and network-client behavior.
Decision evidence
public snapshot- The CLI installs bundled skills into Claude Code and Codex skill directories at startup; installation is enabled by default and can be disabled with an environment setting.
- The inspected Codex usage request sends its saved access token to ChatGPT’s usage endpoint. The cloud credential comparison requests Rudder’s home API and compares account identities.
- The unresolved runtime operations handle text formatting, secret redaction, subprocess output logging, WebSocket headers, and clipboard paste conversion.
- The inspected credential flow did not establish a transfer to an unrelated recipient.
- The skill installation is triggered by running the CLI, not by the npm postinstall hook.
Source & flagged code
24 flagged · loading sourcePackage defines install-time lifecycle scripts.
package.jsonView on unpkgInstall-time source downloads a native archive from a fixed external host without transport verification, extracts it, and installs an executable payload.
package.jsonView on unpkg · L1Findings
Affected versions and remediation
This report applies to @exla/rudder@3.9.5.
See all versions of @exla/rudder for other recorded verdicts.
- Review the evidence and your use of @exla/rudder@3.9.5 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.