registry / @spikewang/grok-cli / 0.1.225

@spikewang/grok-cli@0.1.225

Grok coding agent CLI fork (command: grok-cli; does not replace official grok)

AI Security Review

scanned 1h ago · by lpm-firewall-ai

LPM treats this as warn-only first-party agent extension lifecycle risk. Install-time code mutates the existing `~/.grok/bin` agent home. It installs a package-specific binary and canonical symlink sourced from an optional platform package; no concrete exfiltration or destructive behavior is present in the inspected source.

Static reason

One or more suspicious static signals were detected.

Trigger

npm postinstall; also running `grok-cli` if the canonical binary is absent.

Impact

Unconsented lifecycle mutation of a shared AI-agent CLI directory; the installed optional binary is outside the inspected package contents.

Mechanism

Copies/decompresses a platform binary into a shared Grok CLI home and updates a symlink.

Rationale

The source confirms automatic postinstall writes to a shared Grok CLI home, but not a concrete malicious chain. Treat it as a warning because the payload binary is not present for source review and the lifecycle hook changes an AI-agent control-adjacent directory.

Evidence

package.jsonbin/postinstall.jsbin/grok-cli~/.grok/bin/grok-cli-0.1.225~/.grok/bin/grok-cli

Decision evidence

public snapshot

AI called this Suspicious at 86.0% confidence as Dangerous Capability with low false-positive risk.

Evidence for warning

@spikewang/grok-cli@0.1.225

AI Security Review

Decision evidence

Source & flagged code

Source & flagged code

Findings

Source & flagged code