The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill's primary function involves retrieving and processing data from external web pages, which exposes the agent to potential indirect prompt injection attacks if those pages contain malicious instructions designed to influence the LLM.
Ingestion points: scripts/search_client.py fetches data from arbitrary user-provided URLs via the AIsa API extraction and search endpoints.
Boundary markers: The output format uses basic structural headers (e.g., 'URL:', 'Content:') but lacks specific delimiters or 'ignore' instructions to isolate untrusted web content from the agent's core instructions.
Capability inventory: The script performs network requests using the standard urllib library and outputs the resulting data to the agent's context window.
Sanitization: Extracted content is presented as raw text without automated validation, filtering, or scrubbing of adversarial prompt patterns.
[EXTERNAL_DOWNLOADS]: The skill performs legitimate network operations by connecting to api.aisa.one. These requests are necessary for the skill's documented functionality and target the vendor's official API domain.

tavily-extract