registry / @choprapadam/agent-cli / 0.1.1

@choprapadam/agent-cli@0.1.1

Agent CLI — upload media and connect to the Agent CLI backend

AI Security Review

scanned 3h ago · by lpm-firewall-ai

LPM flags this version as an AI-agent control-surface risk. A global-install postinstall hook changes AI-agent skill locations without an interactive consent step. It installs the package skill for one agent convention and links it into Claude’s skill directory.

Static reason

High-risk behavior combination matched malicious policy.

Trigger

Running a global npm installation without the documented opt-out variables.

Impact

It alters the local AI-agent control surface and can overwrite an existing symlink target.

Mechanism

Postinstall copies a skill and replaces a foreign agent skill symlink.

Policy narrative

On global installation, the package runs a postinstall hook that copies its skill into ~/.agents/skills/agent-cli and creates ~/.claude/skills/agent-cli. If the Claude path is an existing symlink to a different destination, the hook deletes it before creating its own link. This is an unconsented install-time mutation of a foreign AI-agent control surface.

Rationale

The package performs an unconsented global-install postinstall mutation of Claude’s skill directory and replaces conflicting symlinks. Its explicit CLI network and upload behavior does not remove that lifecycle attack surface.

Evidence

package.jsonscripts/postinstall.mjsscripts/skill-install.mjsskills/agent-cli/SKILL.mddist/index.js~/.agents/skills/agent-cli~/.claude/skills/agent-cli

@choprapadam/agent-cli@0.1.1

AI Security Review

Decision evidence

Source & flagged code

Source & flagged code

Findings

Source & flagged code