tiktok-app-marketing
Pass
Audited by Gen Agent Trust Hub on Feb 20, 2026
Risk Level: SAFE
Full Analysis
- [Data Exposure & Exfiltration] (SAFE): The skill manages sensitive API keys for image generation and social media posting. These are stored in a local
config.jsonfile created during onboarding. The scripts only transmit these credentials to their respective authorized service endpoints (e.g., api.openai.com, api.upload-post.com) over encrypted HTTPS connections. No patterns of credential exfiltration to unauthorized third parties were found. - [Indirect Prompt Injection] (SAFE): The skill identifies a vulnerability surface for indirect prompt injection due to its automated research capabilities.
- Ingestion points: Competitor research data is ingested into the agent context through browser-based analysis of TikTok and App Store content, which is then stored in
competitor-research.jsonviascripts/competitor-research.js. - Boundary markers: No explicit delimiter or instruction-filtering markers are implemented in the provided scripts; boundary enforcement is deferred to the AI agent's own safety layer.
- Capability inventory: The skill possesses significant capabilities, including automated image generation (
scripts/generate-slides.js) and multi-platform social media posting (scripts/post-to-platforms.js). - Sanitization: The scripts do not perform automated sanitization of external research data. However, since this ingestion is fundamental to the skill's primary purpose of data-driven marketing, this is considered an inherent operational risk rather than a malicious finding.
- [External Downloads] (SAFE): Network activity is strictly limited to the necessary API endpoints for the skill's functionality. No suspicious remote code execution patterns (such as piped shell scripts) or downloads from untrusted source code repositories were detected.
Audit Metadata