The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and analyze external academic papers and literature, which creates a surface for indirect prompt injection if those documents contain malicious hidden instructions.
Ingestion points: The skill utilizes read_file and search_project tools to process external research materials as described in the 'Paper Analysis Framework'.
Boundary markers: The instructions do not establish explicit delimiters or directives for the agent to ignore instructions embedded within the research papers.
Capability inventory: The skill is granted access to read_file, search_project, and write_file tools, allowing it to both read untrusted data and write artifacts to the workspace.
Sanitization: There is no evidence of sanitization or validation protocols for the external content before it is processed by the agent.

academic-researcher