The Agent Skills Directory

[SAFE]: The skill serves as a structured prompt for conversational exploration and tradeoff analysis. It contains no executable scripts, shell commands, or remote resource fetches.
[SAFE]: The frontmatter configuration disable-model-invocation: true effectively restricts the agent from calling external tools, which significantly limits the potential attack surface.
[SAFE]: The internal rules specifically forbid writing to the filesystem (.scratch/), preventing common persistence and data leakage vectors.
[SAFE]: No patterns of prompt injection, obfuscation, or sensitive data access were detected.

explore