The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted data from URLs or text input to perform its audit, creating a surface for instructions embedded in that data to influence agent behavior.
Ingestion points: Untrusted content is ingested via the content parameter (which can be a URL or raw text) and the competitor_urls list as defined in the Input Schema and workflow steps.
Boundary markers: The instructions lack specific boundary markers or delimiters to isolate untrusted external content from the skill's logic.
Capability inventory: The agent uses text generation and is instructed to use web_search or web_browse tools to fetch competitor data.
Sanitization: There are no instructions for sanitizing, escaping, or filtering the external content before it is processed by the agent.
[EXTERNAL_DOWNLOADS]: The skill is designed to fetch and analyze external content from user-specified URLs and competitor websites.
Evidence: The Input Schema allows for URLs in the content field and a list of competitor_urls. The workflow explicitly mentions fetching content if a URL is provided and using web_search/web_browse for competitive comparisons.

seo-audit