The 22 skills · Free skill · Free, and in both paid packs

AGENT SKILL

honest-dataviz: what it makes your agent do, and what it refuses.

honest-dataviz is the rulebook your agent reads before it draws anything. It picks the form from the job the data is doing, caps the type scale, and spends colour on data and never on chrome. A figure the build could not verify is shown and labelled, never quietly dropped.

The job honest-dataviz does

The skill turns make it look good into decisions with reasons. It starts one step earlier than a charting library: one current value goes to a stat tile, a ratio against a limit goes to a meter, and more than about seven categories that all matter go to a table. Once a chart is right, the encoding follows the job. Magnitude is one hue running light to dark, identity is two or three direct labelled hues, and the green and red pair is reserved for status. Typography is capped at three sizes per view with the weights pulled back, and the mark specs are written down, so the result is right by construction rather than by taste.

What honest-dataviz refuses

A skill is worth what it stops as much as what it produces. These are the refusals written into honest-dataviz, taken from the rules and limits in its own file.

• ✓Colour on chrome. Axes, grids, panels and labels stay neutral. If a colour appears on the page, it means something.
• ✓Green and red as a chart palette. The pair is measured below safe colour blind separation, so it is legal only when position, a glyph or a word carries the polarity too.
• ✓The anti-pattern catalogue. Dual axes, rainbow categoricals, truncated bar axes, a number on every point, pies past two or three slices. If the chart matches an entry, it is wrong.
• ✓Hiding doubt. An unverified figure renders muted, flagged, and with its reason findable, because hiding doubt is the one thing a reader never forgives.

10 lines from the honest-dataviz skill file, verbatim

From skills/honest-dataviz/SKILL.md, unedited. Kymira skills are plain text you can read in full before you install them, which is the point of shipping a method rather than a binary.

- **Colour is spent on data, never on chrome.** Axes, grids, panels and
  labels are neutral. If a colour appears, it means something.
- **Semantic pair**: green = up/good/verified, red = down/fails — and
  NOTHING else ever borrows those two. A red that sometimes means
  "category B" destroys the one thing colour was teaching the eye.
- **The semantic pair is a STATUS pair, not a chart palette.** Measured:
  a standard green/red pair fails colour-blind separation (deutan ΔE ≈ 5,
  where ≥ 8 is safe). Legal only when polarity is ALSO carried by
  position (diverging around zero), glyphs (▲ ▼ ✓), or words — colour
  reinforces, never carries alone.

The failure modes honest-dataviz addresses

These entries in the catalogue of 51 observed failure modes name this skill as what addresses them. Every one was seen in a real build or in our own audits, and none was invented.

The tier honest-dataviz ships in

Both paid tiers are a one-time purchase with lifetime updates, and all sales are final: the files are plain text you keep the moment they arrive, so judge the quality on the free skill and the published doctrine first. The pricing section states exactly what each tier contains.

Installing honest-dataviz

A skill is a folder with a SKILL.md file in it. Installing one is copying that folder where your agent looks for skills, with no build step and no dependencies.

mkdir -p ~/.claude/skills
cp -R honest-dataviz ~/.claude/skills/

The free skill is public: download it from the GitHub repository or from the get page, no email asked. Skills are read at startup, so start a new session and ask your agent which skills it has available.

Not for

Last updated 2026-08-29 · Dylan, founder · one of 22 packaged skills, each one a method written down, not a product tour.