registry / @social-data/social-neuron-subscription / 1.0.4

@social-data/social-neuron-subscription@1.0.4

Social Neuron subscription 能力 CLI —— 一能力一命令(social-neuron-subscription),薄客户端直达 /subscription/{op}。

AI Security Review

scanned 7d ago · by lpm-firewall-ai

LPM flags this version as an AI-agent control-surface risk. Installation automatically writes agent instructions into the foreign OpenClaw skills directory. Existing files in the matching skill directory can be overwritten.

Static reason

One or more suspicious static signals were detected.

Trigger

npm postinstall runs node bin/install-skills.js.

Impact

Changes the instructions and capabilities available to an external AI agent.

Mechanism

The installer copies bundled skill files to OPENCLAW_SKILLS_DIR or the user's .openclaw/skills directory without requesting consent.

Policy narrative

The postinstall hook immediately invokes an installer that copies the bundled subscription skill into OpenClaw's user-level skills directory. The copy overwrites matching files without an explicit setup command or consent gate, modifying a foreign agent's instruction surface during package installation.

AI rationale

The source establishes automatic, unconsented postinstall mutation of a foreign AI-agent control surface, which meets the blocking policy. The limited skill scope and absence of exfiltration do not remove that installation behavior.

Evidence

package.jsonbin/install-skills.jsskills/social-neuron-subscription/SKILL.md~/.openclaw/skills/social-neuron-subscription/SKILL.md

Decision evidence

public snapshot

with false-positive risk.

@social-data/social-neuron-subscription@1.0.4

AI Security Review

Decision evidence

Source & flagged code

Source & flagged code

Findings

Affected versions and remediation

Source & flagged code