The Agent Skills Directory

[COMMAND_EXECUTION]: The skill instructs the agent to execute a local script located at scripts/generate_emergency_card.py to generate various HTML formats of the emergency card. Since the contents of this Python script are not included in the provided skill files, its safety and behavior cannot be verified, representing a risk for arbitrary code execution if the script is tampered with or contains vulnerabilities.
[PROMPT_INJECTION]: The skill is susceptible to Indirect Prompt Injection (Category 8) because it ingests and processes data from multiple external and potentially untrusted sources such as medical discharge summaries and surgical records.
Ingestion points: Reads data from files located in data/出院小结/**/*.json, data/手术记录/**/*.json, and several chronic disease tracker files.
Boundary markers: The instructions lack explicit delimiters or warnings to the model to ignore any embedded instructions within these medical records.
Capability inventory: The skill has the capability to read any local file (readFile, glob), write files (saveFile), and execute shell commands (python scripts/...).
Sanitization: There is no evidence of data validation, escaping, or sanitization performed on the content extracted from the medical JSON files before it is processed or used in formatting.

emergency-card