direct-response-copy

Pass

Audited by Gen Agent Trust Hub on Nov 17, 2026

Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill executes a local Python script (brand_context.py) located in a sibling directory (../start-here/scripts/) to load brand-specific context during the initialization phase if brand files are present.
  • [INDIRECT_PROMPT_INJECTION]: The skill possesses an attack surface for indirect prompt injection because it is designed to scrape and process content from external product pages, competitor websites, and review platforms (e.g., Amazon, Reddit, Yelp) to extract customer language.
  • Ingestion points: Web scraping operations in Phase 0 (Step 2) using tools like Firecrawl and Exa.
  • Boundary markers: No explicit delimiters are specified for the untrusted external content within the instructions.
  • Capability inventory: The skill can execute local Python scripts and perform file writes to the ./campaigns/ and ./brand/ directories.
  • Sanitization: The skill employs a 'claim gate' mechanism that requires all output to be verified against a fact sheet, providing a manual or logical validation step that mitigates risks of malicious instructions being followed.
Audit Metadata
Risk Level
SAFE
Analyzed
Nov 17, 2026, 04:01 PM