The Agent Skills Directory

[PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection because it ingests and processes untrusted data from PDF files.
Ingestion points: Content is read from PDF files in the literature/ directory via the Read tool (SKILL.md).
Boundary markers: The instructions do not define boundary markers or delimiters to isolate the PDF content from the agent's instructions.
Capability inventory: The skill utilizes the Read, Glob, and Grep tools.
Sanitization: No sanitization or content validation is performed on the data extracted from the PDF before it is summarized or quoted.

read