The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and parse untrusted data from external supplier price files (xlsx, csv, 1C exports). This presents an attack surface where malicious instructions could be embedded in the data to influence the agent's behavior. The risk is mitigated by the skill's explicit instructions to present matched lists to the user and require manual confirmation before executing any marketplace updates.
Ingestion points: Price files in any format (xlsx, csv, 1C).
Boundary markers: None explicitly defined for the file parsing step, but user review is mandated.
Capability inventory: File writing and marketplace record updates via wb_content_cards_upload, ozon_prices_update, ym_update_prices, and others.
Sanitization: Relies on human review and the confirm_write mechanism before state-changing actions.
[EXTERNAL_DOWNLOADS]: The documentation suggests installing the marketplaces-mcp-ru server using uvx if it is not already present in the environment. This involves downloading and executing external code from a public package registry to enable the skill's core functionality.

dropshipping-intake