What problem does it solve? Dense relational, proportional, or sequential information is hard to hold in mind as prose, and rendered charts do not survive copy-paste into terminals, chat transcripts, or plain-text documents. This Skill turns such material into text figures that stay readable in any monospace context. ## Core Features & Use Cases - Seven figure shapes: bar, spine, tree, lane, fork, matrix, and small-multiple, each mapped to a specific kind of information such as magnitude, sequence, containment, or branching outcomes. - Strict output contract: printable ASCII only, 80-column hard limit with a 64-column target, computed bar proportions, largest-remainder percentage rounding, and a one-line receipt summarizing what the figure carries. - Bundled mechanical checker: a Node.js script validates fence width, ASCII glyphs, bar-to-label proportionality, and percentage sums before a figure ships. - Use Case: Given a household water-usage breakdown, produce a fenced bar figure with aligned labels, exact per-litre scale, percentages summing to 100, and a caption stating the finding. ## Quick Start Ask the agent to draw this data as a terminal-safe ASCII diagram using the human-draw skill.