The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill ingests data from external web pages on permutive.com, creating a potential vulnerability to indirect prompt injection where malicious instructions embedded in the site's content could influence the agent's behavior.\n
Ingestion points: The tools permutive_com__read_page, permutive_com__get_search, and permutive_com__get_resources are designed to bring external text, links, and titles from permutive.com into the agent's context.\n
Boundary markers: The provided instructions do not include specific delimiters or warning messages to the agent regarding the potential for untrusted content within the fetched data.\n
Capability inventory: The agent has access to network capabilities via the Unbrowse API and can execute broader tasks using tools like unbrowse.run and unbrowse.browse.act.\n
Sanitization: No mention is made of sanitizing, escaping, or validating the ingested content before the agent processes it.

permutive