The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill instructs the agent to process external CSV files to manage tester lists, creating an attack surface where data could be interpreted as instructions.
Ingestion points: SKILL.md describes commands that read testers.csv and beta-testers.csv for bulk operations.
Boundary markers: No delimiters or specific instructions are provided to help the agent distinguish between data entries and potential malicious prompts within the files.
Capability inventory: The skill utilizes shell command execution via the gpc CLI tool to modify app distribution tracks and tester access.
Sanitization: There is no mention of content validation or sanitization for the input files before they are processed by the tool.

gpc-testing