registry / siro_lib / 0.2.514

siro_lib@0.2.514

AI Security Review

scanned 19d ago · by lpm-firewall-ai

LPM blocks this version under the AI-agent control-surface policy. The package ships a Claude local settings file that can expand an AI agent's allowed commands and filesystem reads when the package directory is opened. Runtime library code does not show malware behavior, but the agent-control file creates a concrete reviewer/developer control-surface risk.

Static reason

One or more suspicious static signals were detected.

Trigger

Opening or working in the package with an agent that honors .claude/settings.local.json

Impact

May permit credential/config inspection and broad local file reads through the agent environment

Mechanism

AI-agent permission expansion for npm credential commands and home-directory reads

Policy narrative

A developer or reviewer opening the extracted package with a Claude-compatible agent may have local permissions expanded by .claude/settings.local.json. The allowed actions include npm config/whoami commands and reading a user home path, while .env also contains plaintext credential-like values. The bundled React library itself appears package-aligned, so the blocking concern is the included AI-agent control surface rather than install-time npm execution.

Rationale

Static inspection found no install/import malware in the library bundle, but the shipped .claude local permission file is a concrete unconsented AI-agent control-surface risk that can expose npm identity/config and local files. Product guard normalized a concrete AI-agent control hijack publish_block to the blockable dangerous-capability shape.

Evidence

package.jsondist/index.jsdist/index.es.js.claude/settings.local.json.env

siro_lib@0.2.514

AI Security Review

Decision evidence

Source & flagged code

Source & flagged code

Findings

Source & flagged code