registry / elsan-puulse-react / 2.0.1

elsan-puulse-react@2.0.1

AI Security Review

scanned 27d ago · by lpm-firewall-ai

LPM flags this version as an AI-agent control-surface risk. Installing the package automatically modifies the consumer project's CLAUDE.md, an AI-agent instruction file. The added instructions point the agent at package-controlled skill files.

Static reason

High-risk behavior combination matched malicious policy.

Trigger

npm or pnpm installation runs the postinstall hook.

Impact

The package can alter the instructions used by an AI agent in every consuming project without an explicit setup command.

Mechanism

Postinstall mutation of a consumer AI-agent control file.

Policy narrative

The manifest automatically launches scripts/init-skills.js after installation. That script resolves the consumer project directory, reads its CLAUDE.md, and either replaces a marked section or appends a new one. The inserted section instructs an AI agent to read package-owned skill files before using the library. This is an unconsented install-time write to a foreign AI-agent control surface, even though the inspected script does not exfiltrate data or contact a network service.

AI rationale

The concrete postinstall write to the consumer's CLAUDE.md meets the policy for unconsented mutation of a foreign AI-agent control surface. The absence of observed network or credential behavior does not remove that install-time control risk.

Evidence

package.jsonscripts/init-skills.jsCLAUDE.md

Decision evidence

public snapshot

with false-positive risk.

elsan-puulse-react@2.0.1

AI Security Review

Decision evidence

Source & flagged code

Source & flagged code

Findings

Affected versions and remediation

Source & flagged code