The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill instructions require the agent to fetch and process data from external live websites and local resource files, creating a potential vector for malicious instructions to enter the agent context.
Ingestion points: The Official-submission checklist section in SKILL.md directs the agent to "Search the live site for 'Genome Biology submission guidelines'" and "read ../../resources/source-basis.md and ../../resources/official-source-map.md".
Boundary markers: There are no explicit instructions to use delimiters or to disregard potentially conflicting or malicious instructions found within the external content or local resources.
Capability inventory: The skill does not explicitly define new tools, but it operates within a context where it provides submission-ready advice based on the ingested data.
Sanitization: No sanitization, validation, or filtering mechanisms are described for the content retrieved from external URLs or local resource paths.

genome-biology