Back to skill

Security audit

Live Product Lineup

Security checks for vulnerabilities and agentic risk

Overview

The skill is a normal live-commerce planning guide, but its main instruction file is malformed in a way that prevents reliable review or loading.

Do not install this package as-is until the publisher replaces SKILL.md with clean, human-readable UTF-8 Markdown and it is re-reviewed. If using a repaired version, verify stock counts, discounts, giveaways, and platform rules before generating live scripts, and avoid prompts that ask viewers for precise locations in public chat.

Vulnerability Patterns
  • Insecure Skill Coding PracticesFinds exploitable flaws such as hardcoded secrets or command injection
  • Skill Instruction HijackingAlters the agent's session goals or safety constraints when the skill loads
  • Agent Memory PoisoningWrites attacker-controlled rules into memory that affect later sessions
  • Remote Payload Retrieval and ExecutionFetches external code whose behavior can change after review
  • Embedded Malicious CodeShips malicious scripts inside the skill and executes them locally
Findings (1)

T09 · Insecure Skill Coding Practices

Note
Location
SKILL.md:129
Finding

Malformed and Obfuscated Content in the Skill Definition

Content
View full analysis
Remediation
View remediation
Vulnerability Patterns
  • MCP Tool PoisoningHidden Instructions, Unicode Deception, Parameter Description Injection
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
Findings (4)

Ae3

High
Category
analysis-evasion
Confidence
98% confidence
Finding

The file contains embedded NUL bytes and corrupted binary-looking text inside what should be plain Markdown. In an agent skill, this is dangerous because it can hide instructions from reviewers, break downstream parsers, or create parser-differential behavior where humans and tooling see different content. The abrupt corruption appearing mid-sentence strongly increases suspicion beyond an accidental formatting issue.

Content

No source excerpt is available for this finding.

Ae4

Medium
Category
analysis-evasion
Confidence
80% confidence
Finding

Suspicious Unicode normalization or mixed-script content can be used to disguise text, evade review, or cause different renderings across editors and analysis tools. In this file, that risk is amplified by the confirmed corruption later in the document, making it more plausible that nonstandard encoding or obfuscation is present rather than harmless typography alone.

Content

No source excerpt is available for this finding.

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
89% confidence
Finding

The skill operationalizes urgency tactics, countdowns, giveaways, and stock-based persuasion without guardrails requiring truthful claims, platform-rule compliance, or consumer-protection checks. In context, this can directly enable deceptive or noncompliant selling behavior by encouraging manipulative scarcity messaging and promotional mechanics that may be false, unlawful, or policy-violating if used as written.

Content

No source excerpt is available for this finding.

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
94% confidence
Finding

The script explicitly prompts viewers to share their location in a public live chat without any privacy warning, minimization guidance, or safer alternative. In a TikTok Live context, this can encourage unnecessary disclosure of personal information to the host, platform, and other viewers, increasing privacy and targeting risks.

Content

No source excerpt is available for this finding.