The Agent Skills Directory

[COMMAND_EXECUTION]: The skill utilizes common development tools including grep for code searching, git for inspecting version history and diffs, and test runners like npm and pytest for reproduction. These operations are restricted to the local environment and the project source code.- [SAFE]: The instructions explicitly direct the agent to sanitize error messages by removing hostnames, file paths, and internal data before searching for solutions externally, which protects sensitive information.- [SAFE]: The skill follows a logical and constrained workflow for debugging without attempting to bypass safety filters, access unauthorized credentials, or establish persistence.

investigate