The Agent Skills Directory

Interactive Debugging Capabilities: The skill provides a workflow for driving a live debugger, which includes tools for starting processes and executing code (start_debugging, evaluate_expression). These capabilities are used to verify runtime behavior and are standard for debugging environments.- Runtime State Inspection: The instructions guide the agent to use tools like get_variables_values to inspect local and global application state. This provides the agent with deep visibility into the program's execution to identify the root cause of issues.- Support for Multiple Environments: The skill includes detailed reference guides for Python, JavaScript, Java, C#, C++, and Go, recommending official or well-known extensions and debuggers (e.g., Delve, GDB, LLDB) to ensure a standard and reliable debugging experience.- Indirect Data Ingestion: As a debugging tool, the agent processes information directly from the source code and runtime environment (e.g., file paths, variable values, and expressions).
Ingestion points: Variable values, call stacks, and source code accessed via the debugger tools.
Boundary markers: Not explicitly defined in the workflow.
Capability inventory: Includes tool-driven process execution (start_debugging), expression evaluation (evaluate_expression), and state inspection (get_variables_values).
Sanitization: The skill relies on the underlying debugger's inspection capabilities without additional filtering. This surface is necessary for the skill's purpose, but users should be aware that the agent's reasoning is based on the data provided by the debugged environment.

debug-live