The Agent Skills Directory

[COMMAND_EXECUTION]: The skill documents the use of cargo run and python3 to execute benchmarking and analysis tools located within the perf/memory directory of the repository. This is expected behavior for a development and performance analysis skill.
[INDIRECT_PROMPT_INJECTION]: The skill describes a workflow where the agent analyzes heap profile data (dhat-heap.json) using a Python script. While this involves processing data generated at runtime, the data source (heap profiler) and processing script are internal tools used for performance auditing, presenting a minimal and standard attack surface for a development environment.

memory-benchmark