The Agent Skills Directory

[COMMAND_EXECUTION]: The script scripts/literature_review.py executes sibling scripts like scholar_search.py and arxiv_search.py using subprocess.run. This is used to leverage specialized searching skills within the local environment.\n- [PROMPT_INJECTION]: The skill exhibits a surface for indirect prompt injection via academic metadata fetched from external sources.\n
Ingestion points: External data (paper titles, abstracts, TLDRs) from APIs like OpenAlex and Semantic Scholar are written to candidates.json and shortlist.json in the session directory.\n
Boundary markers: The skill does not employ boundary markers or specific instructions to the agent to disregard instructions potentially embedded in the fetched paper data.\n
Capability inventory: The skill possesses capabilities for command execution (via subprocess) and local file system access (creating and writing to session directories).\n
Sanitization: Fetched metadata is directly saved and processed without sanitization or escaping of its content.

literature-review