What problem does it solve? Littleman programs are spatial, tick-driven boards where performance problems hide in blank-cell commutes, FIFO pipe latency, and stalled runners; reading raw source or numeric logs makes it hard to see where ticks are actually spent. This Skill renders an execution snapshot as an image with exact movement and stall heatmaps so optimization decisions are grounded in observed behavior. ## Core Features & Use Cases - Exact heatmap rendering: Replays a chosen tick window one step at a time to produce precise movement and stall heatmaps, plus runner trails, pipe token positions, and per-room activity counts. - Watched-cell diagnostics: Records before/after runner state (position, direction, A, B, backpack) when runners cross selected cells, supporting decisions about conditional turns on shared lanes. - Use Case: While optimizing an LLLM candidate, render the baseline and the modified program at the same case, tick, and window, then compare the images and JSON hot-cell counts to confirm that a layout change removed a cross-bank commute before submitting through the team judge. ## Quick Start Ask the agent to render the current best .man file for the little-little-little-man case "one tick at a time" at tick 100000 with a 512-tick window and inspect the generated heatmap image for optimization opportunities.
