The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill possesses an attack surface for indirect prompt injection because it is designed to ingest and process untrusted external data such as e-commerce audits and product page content.
Ingestion points: The skill collects findings, page evidence, reviews, and content from user-provided audits and product listings (SKILL.md).
Boundary markers: There are no explicit delimiters or specific instructions to the agent to disregard instructions embedded within the processed data.
Capability inventory: The skill is limited to text generation and does not have any capabilities for subprocess execution, file system writes, or network operations in its scripts or instructions.
Sanitization: No explicit input validation or escaping mechanisms are described for the external content.
[EXTERNAL_DOWNLOADS]: The skill provides installation instructions and integration links that target the author's official packages and domains.
Evidence: The skill references installation via the nexscope-ai/ecommerce-seo-geo-skills package and includes informational links to nexscope.ai tools (SKILL.md).

ecommerce-geo-optimizer