The Agent Skills Directory

[SAFE]: The skill's primary function is educational and investigative, using the Socratic method to guide users toward self-discovery of bugs rather than providing direct solutions.
[COMMAND_EXECUTION]: While the skill includes Bash in its allowed-tools list, the instructions focus on information gathering via Read, Grep, and Glob to inform the agent's questioning. There are no instructions to execute arbitrary or dangerous shell commands.
[DATA_EXPOSURE]: The skill reads relevant code to formulate precise questions. This file access is restricted to the workspace using standard tools and does not target sensitive system configuration files or credentials.
[PROMPT_INJECTION]: The skill instructions do not contain attempts to override safety guidelines or bypass system constraints. It defines its own behavioral constraints (e.g., question-only mode) but includes legitimate protocols for switching to direct assistance modes like systematic-debugging if the user requests it.

socratic-debugging