The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted data from external sources during the product research phase.
Ingestion points: web_search results from external sites (G2, Capterra, TrustRadius) as defined in Step 2 of the workflow in SKILL.md.
Boundary markers: Absent. The instructions lack specific delimiters or warnings to the agent to treat external search results as untrusted or to ignore embedded instructions.
Capability inventory: The skill is configured to use the web_search tool and generates markdown text output. It does not appear to possess file-writing, network-sending (beyond the tool), or command-execution capabilities.
Sanitization: Absent. Data retrieved from external web searches is interpolated directly into the article generation process without evident validation, filtering, or escaping.

listicle-generator