The Agent Skills Directory

[REMOTE_CODE_EXECUTION]: The skill executes local Python scripts to interact with the LinkFox gateway and manage authentication. All network communications are directed to vendor-owned subdomains of linkfox.com (such as tool-gateway.linkfox.com and api.linkfox.com), which is consistent with the skill's primary function and author identity.
[INDIRECT_PROMPT_INJECTION]: The skill processes data from Mercado Libre, which presents an inherent surface for indirect prompt injection. This risk is inherent to the skill's purpose and is managed through the implementation of local data persistence and summarization logic.
Ingestion points: External data is ingested via the unified LinkFox tool gateway in scripts/linkfox_lanjing_mercado_product_selection.py from tools such as itemSearch and reviewSearch.
Boundary markers: The instructions guide the agent to summarize results and prioritize reading from persisted JSON files, avoiding full raw data interpolation into the prompt context.
Capability inventory: The skill's scripts possess capabilities for network requests and local file writing to the ./linkfox directory.
Sanitization: Ingested data is stored in structured JSON files, enabling the use of precise parsing tools like jq to isolate specific fields and minimize the risk of executing embedded malicious instructions.

linkfox-lanjing-mercado-product-selection