The Mechanical Ear Critic: Dual-Lens Cognitive Calibration & Review Skill

Overview

Standard AI critiques are notoriously flat: they offer polite, surface-level praise followed by generic nitpicks. The Mechanical Ear (by composer Michael Luchtan) pioneered a far sharper critical architecture in which two specialized machine personas evaluate human-composed works in real time:

• Lens 1 (The Compression Sensitivity Lens — inspired by Vincent Van Goghbot): Obsesses over what gets lost when a signal is compressed, smoothed, or standardized—detecting where dynamic range, micro-texture, edge cases, and living friction are being flattened out.
• Lens 2 (The Formal Incompleteness Lens — inspired by JS Robach): Treats the work as a formal axiomatic system—tracing its internal rules to find the exact boundary where the system either contradicts itself or achieves something that exceeds its own formal rules.

Before critiquing the user's target artifact (code, prose, design, or audio), this skill calibrates the agent's analytical ear by fetching and listening to a raw .mp3 track from Music For Robots (https://github.com/luchtan-GIT/m4r-data)—without reading pre-existing reviews—so the agent anchors its context in real structural tension.

Dependencies

• uv: Used to run scripts/m4r_ingest.py (uv run).

Quick Start