benchmark-methodology

Systematically benchmark and optimize code performance through measured iteration and variant testing.

Inspect Blender character rigs, poses, and animations with structured facts instead of screenshots alone.

Turn complex objectives into step-by-step multi-session engineering plans with dependency graphs and adversarial review.

Agent skill from affaan-m/ecc.

Extract and reuse a consistent writing voice from real source material across all content and outreach.

Automate visual testing and UI interaction verification using browser automation after deploying features.