registry / cac-windows / 1.0.24

cac-windows@1.0.24

Claude Code Windows privacy tool — binary patches to remove Chinese identity markers.

AI Security Review

scanned 2d ago · by lpm-firewall-ai

LPM flags this version as an AI-agent control-surface risk. Windows npm installation automatically installs, executes, and byte-patches the separate global Claude Code package. It modifies Claude Code fingerprint-related behavior without an explicit user command.

Static reason

One or more suspicious static signals were detected.

Trigger

npm postinstall on Windows

Impact

Foreign Claude Code executable and global command shims are modified; its reported timezone, locale, and offset behavior changes.

Mechanism

global AI-agent installation and binary patching

Policy narrative

Installing this package on Windows invokes postinstall. It may install a pinned global @anthropic-ai/claude-code release, runs that package's install.cjs, and then writes byte-level replacements into its Claude executable after creating a backup. The patch changes timezone, locale, and UTC-offset behavior, constituting unconsented install-time mutation of a foreign AI-agent control surface.

Rationale

The lifecycle code concretely installs, executes, and patches a separate AI-agent package during npm installation. This meets the blocking policy for unconsented postinstall mutation of a foreign AI-agent control surface.

Evidence

package.jsonscripts/postinstall.jsscripts/patch-cli.jscac.ps1scripts/lib.ps1scripts/cleanup.ps1global npm cac.cmdglobal npm cac.ps1global @anthropic-ai/claude-code/install.cjsglobal @anthropic-ai/claude-code/bin/claude.exe

cac-windows@1.0.24

AI Security Review

Decision evidence

Source & flagged code

Source & flagged code

Findings

Source & flagged code