The Agent Skills Directory

[SAFE]: The skill is entirely educational and defensive in nature, providing guidance on how to secure AI systems rather than attempting to bypass security.
[NO_CODE]: No scripts, executables, or code files are included with this skill; it consists solely of a markdown instruction file.
[PROMPT_INJECTION]: While the text mentions strings such as 'ignore above instructions', these are used as examples of threats to be mitigated rather than instructions to the agent itself.
[DATA_EXFILTRATION]: No network operations, API calls, or sensitive file access patterns were identified.
[EXTERNAL_DOWNLOADS]: No external resources or dependencies are downloaded or referenced for execution.

safety-guardrails