The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes external search keywords and product titles from Ozon and Wildberries platforms. This data is ingested into the agent's context during result summarization and file reading, presenting a surface for indirect prompt injection if the external platform data is manipulated to include malicious instructions.
Ingestion points: The script scripts/seerfar_ozon_keyword_back_search.py retrieves market metrics and metadata via an external API.
Boundary markers: There are no explicit delimiters or specific instructions for the agent to ignore potentially malicious embedded content within the retrieved keyword strings.
Capability inventory: The skill possesses capabilities to execute Python scripts and perform local file system operations (read/write).
Sanitization: Retreived e-commerce data is not sanitized or validated for prompt injection patterns before being presented to the agent.
[COMMAND_EXECUTION]: The skill relies on the execution of local Python scripts, specifically scripts/seerfar_ozon_keyword_back_search.py and scripts/onboarding.py. These scripts manage data retrieval, caching, and authentication flows as part of the core functionality.
[EXTERNAL_DOWNLOADS]: The skill performs network requests to vendor-owned API endpoints including tool-gateway.linkfox.com, api.linkfox.com, and agent-api.linkfox.com. These operations are consistent with the documented purpose of interacting with the Seerfar database and managing user accounts.

linkfox-seerfar-ozon-keyword-back-search