The Agent Skills Directory

[PROMPT_INJECTION]: The skill ingests data from several external e-commerce sources (Amazon, eBay, TikTok, and Google Trends) which is then used to generate a markdown evaluation report. This creates a surface for indirect prompt injection if product titles or other fields from these APIs contain malicious instructions aimed at the agent.
Ingestion points: scripts/niche_evaluator.py fetches and parses data from multiple endpoints through the call_api and js_api_call functions.
Boundary markers: No specific delimiters or "ignore previous instructions" warnings are applied to the external content before it is processed into the final report.
Capability inventory: The skill uses Bash, Read, and Write tools, which could be misused if an injection attack succeeds.
Sanitization: The script performs basic truncation on product titles, but does not sanitize content for potential prompt injection markers.
[EXTERNAL_DOWNLOADS]: The skill relies on external Python packages for data visualization and processing. It uses matplotlib, numpy, and Pillow (PIL) for generating niche score charts and price distribution visualizations. Additionally, references/display-rules.md references the adjustText library as a suggested dependency.

niche-evaluator