Agent Ergonomics
How well 7 tools serve an agent
One reference agent, the same task per category, five trials each. Graded against deterministic ground truth — did the output actually render, and does it hold up when repeated? Where a tool was taught more than one way (official docs vs a skill), the sources are compared head-to-head.
7 of 7 tools · headline = the tool's best measured teaching source
| plantuml via skill | A− | diagrams | +0.35 | +0.62 | +0.58 | +0.38 | +0.48 | +0.49 | 0 | +0.50 |
| mermaid via official docs | A− | diagrams | +0.23 | +0.62 | +0.63 | +0.36 | +0.35 | +0.48 | 0 | · |
| d2 via official docs | A− | diagrams | +0.26 | +0.69 | +0.45 | +0.42 | +0.37 | +0.47 | 0 | · |
| matplotlib via skill | A− | data-visualization | +0.39 | +0.59 | +0.33 | +0.25 | +0.15 | +0.38 | 0 | +0.15 |
| react-email via skill | B− | advertising-creative | +0.34 | +0.45 | +0.35 | +0.41 | +0.20 | +0.39 | 1 | · |
| excalidraw via official MCP | B− | diagrams | -0.18 | +0.25 | +0.19 | -0.12 | +0.05 | +0.05 | 5 | · |
| latex via skill | F | documents | +0.32 | +0.14 | +0.05 | +0.35 | +0.17 | +0.20 | 3 | · |
Grades are anchored in a deterministic checker where one exists (renders? requirements met? reliable when repeated?) — a green matrix cannot rescue output that does not work. Surface columns are −1…+1 instrument readings.
Quality × cost
up-left is better · ◆ = on the frontierEach point is one subject: its grade quality against the tokens an agent spends per successful run. The axes stay separate — a cheap unreliable tool and an expensive reliable one are different answers, not one number.