The Agent Skills Directory

[COMMAND_EXECUTION]: The skill executes a local Python script scripts/generate_ai_report.py during Step 8 to generate interactive HTML reports. Executing local scripts can lead to arbitrary command execution if the script logic is compromised or if inputs are not properly sanitized.
[PROMPT_INJECTION]: The skill exhibits an attack surface for indirect prompt injection because it ingests untrusted data from multiple sources and lacks explicit boundary markers or sanitization logic.
Ingestion points: Reads health data from data/profile.json, data/index.json, data/medications.json, data/allergies.json, and tracker files in data-example/ (fitness, sleep, nutrition, and mental health trackers).
Boundary markers: None identified. There are no instructions to the agent to ignore potential commands embedded within the JSON health records.
Capability inventory: The skill has Read and Write permissions and can execute the scripts/generate_ai_report.py subprocess.
Sanitization: No evidence of data validation, escaping, or filtering of the external JSON content before it is processed by the AI or the report generation script.
[DATA_EXPOSURE]: The skill is designed to read highly sensitive files, including data/medications.json (medication history), data/allergies.json (allergy records), and data/profile.json (personal identity and biometric data). While this is the primary purpose of the skill, the access to these paths represents a significant data exposure risk if the agent's behavior is influenced by malicious instructions.

ai-analyzer