Air Duct Cleaning Iaq Kit

• Outcome: Pass
• Latest audit: Sep 16, 2026 · 8:47 AM UTC
• Version: 1.0.1

Overview

This is a markdown-only marketing prompt kit with no executable behavior, though users should verify local legal rules and edit several compliance-sensitive templates before publishing output.

Install risk is low from a system-security perspective because this skill is a static prompt kit. Before using its output in ads, reviews, SMS, or service pages, confirm current local licensing and advertising requirements, especially outside Nevada, and revise the conditional review asks, referral disclosure text, physician-facing phrasing, and factual contamination claims.

A.I.G

Vulnerability Patterns

• Skill Instruction Hijacking: Alters the agent's session goals or safety constraints when the skill loads
• Agent Memory Poisoning: Writes attacker-controlled rules into memory that affect later sessions
• Remote Payload Retrieval and Execution: Fetches external code whose behavior can change after review
• Embedded Malicious Code: Ships malicious scripts inside the skill and executes them locally
• Unauthorized Access and Privilege Escalation: Obtains permissions beyond the task's legitimate needs

SkillSpector

Vulnerability Patterns

• Prompt Injection: Instruction Override, Hidden Instructions, Exfiltration Commands
• Excessive Agency: Unrestricted Tool Access, Autonomous Decision Making, Scope Creep
• MCP Tool Poisoning: Hidden Instructions, Unicode Deception, Parameter Description Injection
• Data Exfiltration: External Transmission, Env Variable Harvesting, File System Enumeration
• Privilege Escalation: Excessive Permissions, Sudo/Root Execution, Credential Access

Findings (11)

Natural-Language Policy Violations

Medium

• Category: Not specified by scanner
• Confidence: 92% confidence

Finding

The skill repeatedly states Nevada-specific legal and licensing rules as if they are generally applicable, but it does not clearly constrain its use to Nevada or require a locale check before applying those rules. In a compliance-marketing skill, this can cause users in other jurisdictions to generate legally inaccurate content or make incorrect operational decisions, creating regulatory, contractual, or consumer-protection exposure.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium

• Category: Not specified by scanner
• Confidence: 93% confidence

Finding

Clark County and Nevada mold-remediation requirements are presented as default enforcement rules without any visible validation that the user operates in that locality. Because the skill is meant to generate legally compliant marketing, silently applying local rules out of context increases the chance of false compliance claims, improper service restrictions, or omitted obligations in other states.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium

• Category: Not specified by scanner
• Confidence: 95% confidence

Finding

The skill enforces Nevada C-21 HVAC licensing requirements for modifications as though they are the default rule for all users, despite the target market including broad contractor categories and agencies that may operate outside Nevada. In context, this is more dangerous because the skill is positioned as a permanent compliance framework; users may rely on it without independent verification and produce inaccurate legal guidance or business offers.

Content

No source excerpt is available for this finding.

Autonomous Decision Making

Medium

• Category: Excessive Agency
• Confidence: 75% confidence

Finding

Skill enables autonomous high-impact decisions without human-in-the-loop verification. Critical operations (destructive commands, financial transactions, data deletion) should require explicit user confirmation.

Content

Scanner excerpt · prompts/01-seasonal-campaign.md (reported line 107)May include surrounding context.

md

- NADCA member # visible
   - Focus: duct system cleanliness after smoke event, not health claims
   - Honest offer: free visual inspection, no obligation
   - TCPA-compliant contact (no SMS without consent, phone/email only)

---

Natural-Language Policy Violations

Medium

• Category: Not specified by scanner
• Confidence: 91% confidence

Finding

The instruction says 'Generate a complete reputation and referral system' and all embedded templates are specified only in English, with no option for the user to choose another language or locale. This creates a natural-language policy concern when a skill effectively forces a specific language without opt-in.

Content

No source excerpt is available for this finding.

Intent-Code Divergence

Medium

• Category: Not specified by scanner
• Confidence: 98% confidence

Finding

This is a true issue because the skill explicitly says review requests must be sent to all completed-service customers and must not be selectively filtered, yet multiple templates condition the ask on the customer being 'happy' or otherwise positive. That creates a review-gating workflow that can violate FTC endorsement expectations and platform integrity rules, exposing the business to compliance and reputational risk.

Content

No source excerpt is available for this finding.

Missing User Warnings

Medium

• Category: Not specified by scanner
• Confidence: 88% confidence

Finding

The referral acknowledgment SMS reveals that a named third party's duct cleaning 'is done' to the referrer, which discloses service completion information about another customer without any stated consent basis. In a home-services context, this can expose private relationship, address, or service-use information and create privacy and trust harms even if no formal medical data is involved.

Content

No source excerpt is available for this finding.

Intent-Code Divergence

Medium

• Category: Not specified by scanner
• Confidence: 99% confidence

Finding

The office follow-up script asks 'How did everything go?' and only after a positive response says 'Great. If you're happy, we'd appreciate a Google review,' which is classic selective review solicitation. In this skill's compliance-focused context, that contradiction is especially problematic because it operationalizes deceptive review filtering despite the document's own stated prohibition.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium

• Category: Not specified by scanner
• Confidence: 92% confidence

Finding

The skill hard-codes multiple legal and regulatory frameworks, including Nevada-specific licensing rules, as if they universally apply. In a general-purpose skill, this can mislead users in other jurisdictions into relying on incorrect legal/compliance guidance, causing unsafe or noncompliant business actions.

Content

No source excerpt is available for this finding.

Intent-Code Divergence

Medium

• Category: Not specified by scanner
• Confidence: 90% confidence

Finding

The prompt claims to block health-outcome claims, yet it includes medically suggestive positioning such as 'Physician-Referral Friendly' and 'Data for you and your physician.' Even without explicit disease claims, this can steer the model to generate regulated health-adjacent advertising that implies medical utility or clinical endorsement, creating legal/compliance risk and misleading marketing output.

Content

No source excerpt is available for this finding.

Intent-Code Divergence

Low

• Category: Not specified by scanner
• Confidence: 80% confidence

Finding

The ad copy states as fact that wildfire smoke particles are in the duct system, while the surrounding compliance framing says such claims should be tied to documented visual evidence. This inconsistency can cause the model to produce unsubstantiated contamination claims, increasing deception and advertising-compliance risk even if the direct security impact is limited.

Content

No source excerpt is available for this finding.

Static analysis

No suspicious patterns detected.