The Agent Skills Directory

[SAFE]: The skill serves as a template for designing secure systems. It provides high-quality instructions for implementing agent identity, trust scoring, and audit trails.
[COMMAND_EXECUTION]: The provided scripts/run.sh is a standard wrapper script used to execute the skill within its intended repository environment. It does not perform any unexpected or high-risk operations.
[DATA_EXFILTRATION]: No network operations or sensitive data access patterns were identified. The cryptographic keys and JSON schemas included in the documentation are provided as non-functional examples for design purposes.
[PROMPT_INJECTION]: The instructions use standard framing to define the agent's persona and rules. There are no attempts to bypass safety filters or override system instructions in a malicious way.

agentic-identity-trust