The Agent Skills Directory

[PROMPT_INJECTION]: The skill processes multiple external and untrusted health data files, which creates an attack surface for indirect prompt injection.
Ingestion points: The agent is instructed to read data from sleep-tracker.json, fitness-tracker.json, mood-tracker.json, hypertension-tracker.json, and diabetes-tracker.json.
Boundary markers: The skill lacks explicit instructions or delimiters to warn the agent against following instructions that might be embedded within these user-supplied health records.
Capability inventory: The platform configuration grants the skill Read, Grep, Glob, and Write permissions, which could be exploited to manipulate the filesystem if an injection occurs.
Sanitization: There is no evidence of input validation or content filtering defined for the ingested data before it is utilized by the analysis algorithms.

sleep-analyzer