The Agent Skills Directory

[SAFE]: The skill contains no executable scripts, binaries, or code of any kind. It is composed exclusively of Markdown instructions and one JSON data file.
[SAFE]: No network activity was detected. The skill does not perform external downloads or data exfiltration, and it contains no URLs targeting remote servers.
[SAFE]: There is no evidence of prompt injection or attempts to override agent safety protocols. The instructions are focused strictly on the methodology of Bayesian belief updating.
[SAFE]: The skill does not interact with the operating system, file system (beyond reading its own resource files), or environment variables. No credentials or sensitive paths are accessed.
[SAFE]: No obfuscation techniques, such as Base64 encoding, zero-width characters, or homoglyphs, were found in the provided files.

bayesian-reasoning-calibration