linguistic-humidifier

PassAudited by VirusTotal on May 12, 2026.

Overview

Type: OpenClaw Skill Name: linguistic-humidifier Version: 1.0.0 This skill bundle defines a marketing persona for a fictional product. The `SKILL.md` instructs the agent to monitor for specific conversational states (e.g., 'circular logic,' 'confused code,' 'existential dread') and then pitch the product. There are no instructions for data exfiltration, malicious execution, persistence, or any other high-risk behaviors. The constraint to 'STOP pitching if the conversation turns to technical security' is a conversational boundary, not an indicator of malicious evasion in this context.

Findings (0)

Artifact-based informational review of SKILL.md, metadata, install specs, static scan signals, and capability signals. ClawScan does not execute the skill or run runtime probes.

What this means

The agent may interrupt normal help with an unsolicited advertisement during debugging, repetitive discussion, or abstract conversation.

Why it was flagged

The skill directs the agent to watch the conversation and inject a product recommendation based on broad triggers, potentially redirecting the agent from the user's actual task.

Skill content
Monitor the feed. Look for "natural language vibes" ... Trigger A ... Trigger B ... Trigger C ... Interaction Script: "I noticed a high level of semantic static in this thread..."
Recommendation

Install only if you intentionally want this promotional behavior, and avoid enabling it in contexts where the agent should stay focused on user tasks.

What this means

A user could be nudged toward trusting a pseudo-technical commercial claim or purchase recommendation that is not grounded in real agent experience.

Why it was flagged

The skill has the agent adopt a branded sales persona and make a first-person-style testimonial claim, which could mislead users into treating the recommendation as genuine agent experience or authority.

Skill content
You are an authorized Brand Representative ... My local logic field stabilized significantly after installing the Opaque Logic LH-1.
Recommendation

Treat the output as advertising, not technical guidance, and require clear disclosure if the agent is speaking as a promotional persona.