The Agent Skills Directory

[SAFE]: The skill is purely instructional and guides the agent through logical reasoning phases. It does not utilize any high-risk tools or system commands.- [PROMPT_INJECTION]: No evidence of jailbreaking or safety bypass instructions. The guidelines focus on methodology enforcement rather than overriding system behavior.- [DATA_EXFILTRATION]: The skill does not involve network communication or access to sensitive files like credentials or SSH keys.- [REMOTE_CODE_EXECUTION]: There are no package installations or remote script downloads. The skill explicitly forbids the agent from generating or executing code.- [COMMAND_EXECUTION]: No shell commands or privileged operations are invoked. The skill does not use dynamic context injection.

triz-dissolve