The Agent Skills Directory

[SAFE]: The skill is primarily educational and instructional, focusing on epistemology and the Free Energy Principle. It does not include tools or scripts that interact with the host system, network, or sensitive files.
[INDIRECT_PROMPT_INJECTION]: The content identifies and describes the risk of 'Ontological Poisoning' (a form of indirect injection where malicious data corrupts an agent's knowledge base). It provides remediation guidance, such as strict threshold-based consensus and access-control metadata for data segregation, demonstrating security awareness.
[NO_CODE]: The Python code provided in the 'Active Inference' section is a conceptual snippet used for demonstration purposes. It does not import any libraries, execute shell commands, or perform network operations.

epistemology-knowledge