The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill requires the agent to browse and synthesize information from external, third-party websites to extract frameworks and authorities.
Ingestion points: External URLs targeted during research tracks in SKILL.md and references/authorities-and-frameworks.md.
Boundary markers: The instructions lack explicit delimiters or specific warnings to ignore instructions embedded within the browsed content.
Capability inventory: The skill does not possess high-risk capabilities such as arbitrary code execution, network exfiltration of local data, or system-level file access. Its primary output is a markdown document.
Sanitization: There are no specific instructions to sanitize or validate the external content beyond checking if the site is "live" and the "canonical authority".

domain-research