The Agent Skills Directory

[PROMPT_INJECTION]: The skill ingests and processes content from up to 120 externally sourced research papers, which constitutes an indirect prompt injection surface.\n
Ingestion points: Research papers collected via the paper-navigator dependency (SKILL.md).\n
Boundary markers: The instructions lack specific delimiters or instructions to the agent to treat ingested paper content as untrusted or to ignore embedded directives.\n
Capability inventory: The skill utilizes write_file and edit_file to output survey artifacts to the local filesystem (SKILL.md).\n
Sanitization: No explicit validation or filtering mechanisms are described for the external content before processing.

research-survey