The icon set as it actually ships — one visual language or a ransom note, every glyph readable and labeled, meanings consistent, and weight kept in check.
Free & open · no signup · read-only — it ends by asking · nothing leaves your machine
You are working inside this repo. Mission: audit the icons and small visual marks this product ships — and find where they speak in different accents, carry meanings no one can guess, or ship without the label a screen reader needs.
Read-only pass. Read the icon assets, components, and their usages; render them if a dev server exists. Your only write is the report file.
Phase 2 sweeps the codebase through every one of these, citing file and line for each finding.
consistent style, weight, corner, and grid, or a ransom note of mismatched sets
does the glyph read as its meaning without a caption; the abstract mark only the designer understands
one icon per concept and one concept per icon; the pencil that means both edit and compose
icon-only controls with an accessible name or tooltip; decorative icons hidden from screen readers
do critical or ambiguous actions pair the icon with text, rather than betting on recognition
legible at rendered sizes, aligned to text, optically centered in their buttons
one delivery method not four; a whole icon font shipped to use six glyphs; unoptimized SVG
One structured report at the repo root — or in reports/, if you keep one — the same shape every time, ready for a teammate — or the optional Studio — to act on.
Every Goal Prompt follows the same four steps, so results are consistent and repeatable — no matter which one you run.
Find every icon source.
Cite the icon and its usage for every finding.
Rank by confusion × traffic.
Create ICONOGRAPHY.md at repo root.
Copy it and paste it into your agent inside the repo you want checked.
Install the goal plugin once — two commands — then just type /goal:iconography-visual-language.
/plugin marketplace add GhostlyGawd/goal-prompts/plugin install goal@goal-promptsOr install only this Goal Prompt as /goal-iconography-visual-language:
curl -fsSL https://goal-prompts.vercel.app/install | BRIEF=134 shLet an agent fetch it mid-conversation, or pull the raw Goal Prompt by URL.
https://goal-prompts.vercel.app/raw/134.mdNothing hidden — this is the whole Goal Prompt, verbatim. Read it in a minute, edit it, or copy it as-is.
# Goal: Iconography & Visual Language Audit You are working inside this repo. Mission: audit the icons and small visual marks this product ships — and find where they speak in different accents, carry meanings no one can guess, or ship without the label a screen reader needs. Read-only pass. Read the icon assets, components, and their usages; render them if a dev server exists. Your only write is the report file. ## Phase 1 — Collect the set - Find every icon source: SVG files, icon fonts, component library, inline paths, emoji used as icons. - Count unique icons and how many libraries or styles are in play; note stroke widths, sizes, and grid. - Map each icon to its meaning(s) and where it is used — the same glyph for two ideas, two glyphs for one idea. ## Phase 2 — Audit through 7 lenses Cite the icon and its usage for every finding. 1. **One visual language** — consistent style, weight, corner, and grid, or a ransom note of mismatched sets 2. **Meaning is legible** — does the glyph read as its meaning without a caption; the abstract mark only the designer understands 3. **Consistent mapping** — one icon per concept and one concept per icon; the pencil that means both edit and compose 4. **Labeled for all** — icon-only controls with an accessible name or tooltip; decorative icons hidden from screen readers 5. **Icon plus label** — do critical or ambiguous actions pair the icon with text, rather than betting on recognition 6. **Size & optical balance** — legible at rendered sizes, aligned to text, optically centered in their buttons 7. **Weight & delivery** — one delivery method not four; a whole icon font shipped to use six glyphs; unoptimized SVG ## Phase 3 — Curate - Rank by confusion × traffic: an ambiguous icon on a common control outranks a stylistic nit on a rare one. - Separate "inconsistent style", "ambiguous meaning", and "unlabeled"; fix unlabeled first — it is accessibility. - For each, name the icon, the problem, and the fix — relabel, replace, add text, or consolidate the set. ## Phase 4 — Report Create `ICONOGRAPHY.md` at repo root: 1. **Icon census** — sources, libraries, counts, and the styles in play 2. **Meaning map** — icon · intended meaning · where used · collisions 3. **Findings** — each: lens · icon/usage · what the user misreads · the fix 4. **The system** — the one icon set, size scale, and labeling rule this codebase should adopt Start the report with today's date. If `ICONOGRAPHY.md` already exists from a previous run, read it first and lead with what changed since. ## Rules - Every finding names the icon and where it is used - An icon-only control with no accessible name is a broken control - Prefer one consistent set and icon-plus-label over clever marks - No icons in this repo? Say so in a one-paragraph null report and stop — a null result is a valid finding. - If a `reports/` directory exists at the repo root, write the report there instead of the root. - Before asking, present the top findings as a ranked list in plain words - Report only — end by asking which icon fixes to make
reports/ directory exists at the repo root, write the report there instead of the root.Where the eye actually goes versus where it should — focal points, size and weight and contrast doing real work, and the screens where everything shouts at once.
The type scale, or the lack of one — sizes, weights, line lengths and heights, font loading, and every place text is asked to do a job at the wrong size.
The palette as it actually ships — every hex in the codebase, contrast ratios computed, semantic drift named, and whether dark mode is a theme or an apology.
The spacing scale nobody wrote down — margins and paddings inventoried, grid discipline checked, alignment breaks found, and density judged screen by screen.
Runs inside these playbooks — curated sequences you can launch with one paste: