ecommerce-competitor-analyzer

Pass

Audited by Gen Agent Trust Hub on May 2, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: The skill uses official and reputable API endpoints from Google and Olostep to perform its core tasks of data analysis and web scraping.
  • [SAFE]: Secret management is handled via environment variables and local configuration files, following standard security practices to prevent credential exposure.
  • [SAFE]: The implementation of Google Sheets access follows the standard OAuth2 protocol, including secure token handling and clear instructions for users to set up their own credentials.
  • [PROMPT_INJECTION]: The skill processes untrusted product data from Amazon without explicit boundary markers. Evidence chain: 1. Ingestion point: 'scripts/scrape-amazon.js' (product data). 2. Boundary markers: Absent in 'prompts/analysis-prompt-base.md'. 3. Capability inventory: File writes (Markdown reports) and network calls (Google Sheets API). 4. Sanitization: Absent. This finding represents an indirect prompt injection surface which is inherent to the skill's primary purpose of analyzing external web content.
Audit Metadata
Risk Level
SAFE
Analyzed
May 2, 2026, 01:04 PM