The Agent Skills Directory

[PROMPT_INJECTION]: The skill uses authoritative language like 'CRITICAL: Auto-start on load' and 'Do not summarise' to guide the agent's interaction flow. These are standard UI/UX instructions for streamlining agent behavior and do not attempt to bypass safety filters or core system instructions.
[DATA_EXPOSURE]: The instructions reference an optional project file named 'voice.md' to adapt the output tone. This is a localized context reference and does not involve unauthorized access to sensitive files or data exfiltration.
[SAFE]: The skill relies on standard user input gathering and text manipulation. It does not use external dependencies, network operations, or shell commands.

post-formatter