@exla/rudder@3.9.1
⚠ Under reviewA 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 23h ago · by lpm-firewall-aiThe inspected package behavior does not establish a concrete attack: its install hook sets up Jujutsu, and the reviewed credential and agent-related paths align with the product’s cloud and local tooling functions. However, the selected @xterm/headless runtime dependency is loaded by the package and its bundled source remains materially unresolved, so this review cannot certify the package as clean.
Decision evidence
public snapshot- The package declares a runtime dependency on @xterm/headless, and dist/predict.js loads it when the terminal echo predictor is used.
- The dependency’s selected entry is a large bundled file; relevant behavior remains unresolved after the bounded source inspection.
- Cloud credential staging is part of explicit cloud functionality, and keychain access is attempted only when a human is at a terminal.
- The reviewed agent-related locations read local skills, conversation metadata, and model cache data.
- The postinstall hook imports the Jujutsu setup helper and skips setup when requested or running as root.
- Codex usage credentials are sent to the ChatGPT usage endpoint, consistent with the quota lookup.
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@3.9.1.
See all versions of @exla/rudder for other recorded verdicts.
- Review the evidence and your use of @exla/rudder@3.9.1 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.