The Agent Skills Directory

[DYNAMIC_EXECUTION]: The evaluate_expression tool provides the capability to execute arbitrary Java code within the debugged application's runtime environment. The skill documentation includes warnings regarding potential side effects when evaluating expressions that modify application state.
[INDIRECT_PROMPT_INJECTION]: The skill ingests and processes runtime data from the debugged application, creating an attack surface for indirect prompt injection. * Ingestion points: Inspection tools such as get_stack_trace, list_threads, and evaluate_expression retrieve data directly from the active JVM. * Boundary markers: The skill does not define specific delimiters or instructions to ignore potential injection patterns within runtime data. * Capability inventory: The agent has the ability to set breakpoints (add_breakpoint), launch processes (debug_run_configuration), and execute dynamic code (evaluate_expression). * Sanitization: No validation or sanitization of the inspected runtime state is described.
[COMMAND_EXECUTION]: The tool debug_run_configuration is used to initiate the execution of processes defined within the IDE's run configurations.
[EXTERNAL_DOWNLOADS]: The skill instructions provide a link to the vendor's marketplace at amplicode.ru for plugin installation.

java-debug