The Agent Skills Directory

[PROMPT_INJECTION]: The skill ingests untrusted data from external sources, specifically Amazon product reviews through the get_amazon_reviews tool and web search results via ai_search. This creates a potential surface for indirect prompt injection where malicious instructions embedded in processed reviews or webpages could attempt to influence the agent's logic. Ingestion points: Product reviews and web search summaries. Boundary markers: The instructions do not specify the use of delimiters or 'ignore' warnings for these data streams. Capability inventory: The agent has access to market filtering, search, and trend tools. Sanitization: No specific sanitization methods are defined for external data.\n- [EXTERNAL_DOWNLOADS]: The skill is designed to fetch market telemetry, keyword trends, and patent data from the vendor's infrastructure at pangolinfo.com. These interactions are standard for the skill's functionality and target the author's own domain.\n- [NO_CODE]: The skill package contains only instructional markdown and configuration; it does not include any executable scripts, binaries, or local code files.

amazon-product-explorer