The Agent Skills Directory

[SAFE]: The skill communicates exclusively with the official openFDA API at api.fda.gov. There are no unauthorized external connections or signs of data exfiltration.
[SAFE]: API credentials are managed securely via environment variables or optional initialization parameters. The documentation uses standard placeholders and does not contain hardcoded secrets.
[SAFE]: The Python implementation uses standard, reputable libraries like requests and hashlib. No dynamic code execution, obfuscation, or persistence mechanisms were found.
[SAFE]: The local file system is used only for a specific caching directory (fda_cache), which is a standard performance optimization and poses no security risk.
[SAFE]: The skill's behavior is entirely consistent with its stated purpose of regulatory data analysis, and no evidence of prompt injection or metadata poisoning was discovered.

fda-database