linkshell-cli@0.6.4
⚠ Under reviewFollow, steer and approve the coding agents on your computer (Claude Code, Codex, Gemini, Copilot and more) from your phone
AI Security Review
scanned 26d ago · by lpm-firewall-aiInspection found package-aligned screen capture, port discovery, authentication, and upgrade behavior, alongside a runtime agent-settings sweep whose effective scope and caller consent were not established. Relevant static analysis also reports parse errors, so a clean or malicious conclusion is not supported by the inspected source.
Decision evidence
public snapshot- package.json defines a postinstall hook that runs scripts/postinstall.mjs.
- The hook-command implementations enumerate Claude, Codex, and Gemini settings files; the source implementation parses hook entries and writes a file only when its sweep reports a change.
- The Claude SDK code uses the Claude projects directory for project data.
- The screen fallback captures a local screen image; the ports utility enumerates listening ports.
- Login and upgrade contain service-authentication and package-upgrade behavior, while the cited host code handles environment selection.
- The runtime hook sweep writes agent settings files, and the inspected excerpt does not establish the exact filter or caller consent.
- Static analysis reports parse errors in relevant source files, leaving the complete behavior unresolved.
Source & flagged code
10 flaggedPackage defines install-time lifecycle scripts.
package.jsonView on unpkgInstall-time lifecycle script is not statically allowlisted and needs review.
package.jsonView on unpkgPackage source references child process execution.
dist/host/src/ports.jsView on unpkg · L1Findings
Affected versions and remediation
This report applies to linkshell-cli@0.6.4.
See all versions of linkshell-cli for other recorded verdicts.
- Review the evidence and your use of linkshell-cli@0.6.4 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.