The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill references an official documentation link for evidence delivery located at github.com/garrytan/gbrain. As this belongs to the skill's author/vendor, it is considered a legitimate resource reference.
[COMMAND_EXECUTION]: The documentation provides examples of CLI usage for gbrain, such as gbrain recall, gbrain query, gbrain search, and gbrain doctor. These are standard operational commands for the specialized knowledge brain tool and do not involve arbitrary or dangerous command injection.
[REMOTE_CODE_EXECUTION]: No remote code execution or untrusted script downloads were found. All tools are defined as internal MCP tools (recall, search, query, etc.).
[PROMPT_INJECTION]: The skill includes instructions to strictly avoid hallucination and use specific phrases for missing information (e.g., "the brain doesn't have information on X"). These are protective constraints rather than malicious injections.
[DATA_EXFILTRATION]: No evidence of data exfiltration or unauthorized network operations was found. The skill operates on the provided brain context using local or internal MCP tools.

query