The Agent Skills Directory

[PROMPT_INJECTION]: The skill possesses a vulnerability surface for indirect prompt injection through its data persistence mechanism.
Ingestion points: User-provided text strings are stored in fields such as description, notes, and findings within the data/skin-health-tracker.json file.
Boundary markers: The skill instructions do not specify any delimiters or instructions for the agent to ignore or isolate embedded instructions when reading from the local database.
Capability inventory: The skill utilizes Read and Write tools to manage a local database and performs natural language generation when analyzing health status and trend reports.
Sanitization: There is no evidence of text sanitization, validation, or escaping of the user-provided content before it is re-interpolated into the agent's active context during reporting.

skin-health