The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted user input to design and implement neural network architectures, creating a potential vulnerability surface for indirect prompt injection.
Ingestion points: User requests detailing neural network requirements, as shown in the image classification and text generation examples in SKILL.md.
Boundary markers: Absent. The instructions do not define delimiters or specific constraints to prevent the agent from following instructions embedded within user requests.
Capability inventory: The skill uses tools including Bash(cmd:*), Read, Write, Edit, Grep, and Glob to interact with the system and the neural-network-builder plugin.
Sanitization: Absent. There is no mention of input validation or sanitization before requirements are used to generate plugin configurations or execute the build-nn command.

building-neural-networks