The Agent Skills Directory

[SAFE]: The skill is entirely instructional, focusing on improving the agent's reasoning processes (e.g., frame auditing, hypothesis generation, evidence assessment). No executable code, shell commands, or network operations are included.
[SAFE]: The instructions promote defensive behaviors such as "never guess" and "self-report skepticism," which enhance the reliability of the agent's outputs.
[SAFE]: Analysis of the skill metadata and body revealed no indicators of prompt injection, obfuscation, or data exfiltration. References to external files are local to the repository structure.

critical-thinking