The Agent Skills Directory

[PROMPT_INJECTION]: The skill processes untrusted external data within the paper_collection input which presents a surface for indirect prompt injection.
Ingestion points: The paper_collection variable in prompt.md acts as the entry point for external data.
Boundary markers: No specific delimiters or 'ignore instructions' warnings are used to isolate the paper content from the agent's logic.
Capability inventory: The skill currently lacks dangerous capabilities such as file system writes, network requests, or shell command execution, limiting the impact of any potential injection.
Sanitization: No explicit sanitization or validation of the input data is performed.

taxonomy-mapping