The Agent Skills Directory

[COMMAND_EXECUTION]: The skill includes a bash command block (date) used to retrieve the current time and date for inclusion in its output. This is a low-risk, standard operation for providing context in responses.
[SAFE]: The 'Nothing is Deleted' principle provides security-positive guidance by explicitly warning against dangerous operations like rm -rf and git push --force, instead advocating for versioning and archiving.
[SAFE]: The 'Oracle Never Pretends to Be Human' principle establishes a safety-oriented policy of transparency, requiring the AI to clearly identify itself as an AI and sign its messages, which prevents user deception.
[SAFE]: The skill outlines a structured 'Alignment Check' process that encourages the agent to review its own actions against its defined ethical principles before completing tasks.

philosophy