The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill describes a project for building a data processing pipeline that loads and transforms large CSV files. Ingesting untrusted external data creates a surface for indirect prompt injection.
Ingestion points: Project 2 (Data Processing Pipeline) in SKILL.md mentions loading 1GB+ CSV files.
Boundary markers: Not specified in the project description.
Capability inventory: The skill context includes file system access for reading/writing and Python code execution.
Sanitization: No specific sanitization or validation steps are outlined for the data ingestion process.
[DYNAMIC_EXECUTION]: The content introduces Numba's JIT (Just-In-Time) compilation to optimize execution speed by compiling Python code to machine code at runtime.
Evidence: Use of the @jit(nopython=True) decorator in the "High-Performance Computing" section of SKILL.md. This is a standard and expected optimization technique within this skill's domain.
[EXTERNAL_DOWNLOADS]: The skill references several external packages and tools essential for performance work, including NumPy, Numba, and various profiling utilities.
Evidence: Mentions of NumPy, Numba, Cython, py-spy, and Scalene in SKILL.md. These are well-known services and industry-standard tools commonly used for performance analysis and scientific computing.

Python Performance