The Agent Skills Directory

[SAFE]: The skill consists of instructional content that defines a specific persona and workflow for debugging tasks. It does not introduce any malicious behavior.
[PROMPT_INJECTION]: No evidence of instructions designed to bypass safety filters, extract system prompts, or override agent constraints. The 'Core Principle' and 'Natural Debugging Flow' sections serve to guide the agent's reasoning style rather than subvert its behavior.
[EXTERNAL_DOWNLOADS]: No remote URLs or scripts are referenced for download. While the documentation mentions npm install as an example of fixing a missing dependency, this is presented as a standard development scenario and not a hidden or suspicious download.
[COMMAND_EXECUTION]: The skill describes using terminal commands for running tests and verifying fixes (e.g., Bash commands for playtests). These operations are well within the expected scope of a debugging assistant and are not used for unauthorized access or exfiltration.
[DATA_EXFILTRATION]: No patterns associated with data harvesting or sending sensitive information to external domains were detected.

fable-debug