The Agent Skills Directory

[DATA_EXPOSURE] (SAFE): The skill demonstrates how to log LLM requests, responses, and database contents. While this involves sensitive data handling, it is the explicitly stated purpose of the debugging skill and does not involve unauthorized exfiltration.
[EXTERNAL_DOWNLOADS] (SAFE): The skill references the @convex-dev/agent package. This is a standard dependency for the Convex ecosystem and is used for its intended purpose of context inspection.
[COMMAND_EXECUTION] (SAFE): No shell commands or arbitrary code execution patterns were detected; all operations use standard Convex framework APIs.
[PROMPT_INJECTION] (SAFE): No instructions attempting to override agent behavior or bypass safety filters were found.

Convex Agents Debugging