The Agent Skills Directory

[COMMAND_EXECUTION]: The scripts/check.py script executes local Python-based linting and validation tools using list-based subprocess.run arguments. This code execution is benign and restricted to internal development and asset verification tasks.
[INDIRECT_PROMPT_INJECTION]: The skill processes external Python source files and profiler data provided by the user. Potential indirect prompt injection vectors are thoroughly mitigated because input data undergoes rigid structure extraction via Python's built-in abstract syntax tree (AST) parser or rigorous regular expression-based line parsing prior to being interpreted by the model.

performance-profiler