The Agent Skills Directory

[COMMAND_EXECUTION]: The skill provides numerous templates using the Node.js child_process module to execute system commands such as git, ripgrep, find, cat, and jq for codebase introspection and repository management.
[EXTERNAL_DOWNLOADS]: The 'Analysis Mode' reference mod uses curl to interact with the Letta API via the $LETTA_BASE_URL environment variable for agent and conversation introspection.
[DATA_EXFILTRATION]: The skill demonstrates how to read sensitive application metadata, including agent system prompts, memory blocks, and full message history, to generate diagnostic readouts which are then persisted in local state files.
[PROMPT_INJECTION]: The 'Analysis Mode' implementation employs a system reminder technique that instructs the agent to override its standard persona, adopt a clinical monotone, and ignore previous behavioral instructions until a specific exit command is received.

creating-mods