The Agent Skills Directory

[COMMAND_EXECUTION]: The skill requests access to the Bash, Read, Grep, and Glob tools to perform its auditing functions. It utilizes these tools to search for test patterns and analyze source code structure.
[COMMAND_EXECUTION]: The skill uses relative path traversal (../codeprobe/) to load shared configuration files and reference data from the local file system, which is a standard pattern for modular skill architectures.
[PROMPT_INJECTION]: The instructions include specific constraints on how findings should be reported, such as capping the maximum severity at 'Major'. This defines the business logic for the audit rather than attempting to bypass the agent's safety protocols.

codeprobe-testing