The Agent Skills Directory

[PROMPT_INJECTION]: The instructions include directives to maintain the persona across multiple conversation turns ("ACTIVE EVERY RESPONSE", "No revert after many turns"). These are standard for persona-based skills and are balanced by explicit 'Auto-Clarity' rules that require the agent to revert to normal prose for security warnings, irreversible actions, and complex multi-step sequences to prevent user confusion.
[SAFE]: No malicious patterns such as credential harvesting, unauthorized network access, obfuscated code, or dynamic execution were found. The skill operates purely as a set of linguistic constraints on the agent's output.

caveman