The Agent Skills Directory

[NO_CODE]: The skill is entirely declarative, utilizing markdown for instructions and JSON for data structures. It contains no python scripts, javascript files, or other executable code.\n- [PROMPT_INJECTION]: The skill presents an indirect prompt injection surface due to its reliance on local data ingestion without sanitization.\n
Ingestion points: data/index.json and record files in the data/ directory.\n
Boundary markers: Not present. The agent processes medical record content directly without structural isolation.\n
Capability inventory: The skill uses Read and Write tools to interact with the medical data files.\n
Sanitization: No sanitization is performed on the data ingested from files, which could include instructions or malformed paths.\n- [SAFE]: No malicious behavior, obfuscation, or data exfiltration was detected. The skill's functionality is limited to formatting and summarizing medical records as requested.

query-health