Skip to content
Back to skills

Glyphcss Diagrams

ASecurity

Use whenever a reply needs a diagram or chart instead of hand-drawn boxes/arrows/bars or a markdown table: flows, pipeline, architecture, command tree, decisions, dependencies, message sequence (who calls what), state machine/lifecycle, forks and rejoins (branch/merge histories, parallel workstreams, branching stories, revision/experiment lineage), class diagrams, data model/schema/ER/entities, C4/context diagram, mind map/brainstorm, requirements traced to tests; or compare, trend, distribut...

  • 2 stars
  • 0 votes
  • 0 copies
  • 0 views
  • Added October 18, 2026
ai-agentsgonodegit

Works with

  • terminal
  • cli

Security analysis

A100/100

Pro scans all 5 files and shows the line behind each finding

Scanned October 18, 2026

npx -y skills add apresmoi/glyphcss-diagrams --skill glyphcss-diagrams --agent claude-code

Installs into .claude/skills of the current project.

Are you the author of Glyphcss Diagrams?

Add the live security badge to your README. It updates with every re-scan.

Security grade badge for Glyphcss Diagrams
[![Security: A — Skills Directory](https://www.skillsdirectory.com/api/skills/apresmoi-glyphcss-diagrams/badge)](https://www.skillsdirectory.com/skills/apresmoi-glyphcss-diagrams)

More formats (shields.io, HTML) on the badges page. Keep it an A: scan every change in CI with Pro.

Download with Pro
SKILL.md
---
name: glyphcss-diagrams
description: Use whenever a reply needs a diagram or chart instead of hand-drawn boxes/arrows/bars or a markdown table: flows, pipeline, architecture, command tree, decisions, dependencies, message sequence (who calls what), state machine/lifecycle, forks and rejoins (branch/merge histories, parallel workstreams, branching stories, revision/experiment lineage), class diagrams, data model/schema/ER/entities, C4/context diagram, mind map/brainstorm, requirements traced to tests; or compare, trend, distribution, breakdown, scatter, histograms, heatmaps, funnels, Sankey flows, shares of a whole, schedule/roadmap, task boards/kanban, prioritisation/2x2 matrices, multi-criteria scores/radar, budget hierarchies/treemaps, user journeys, bit/byte/packet/header layouts. Triggers include "diagram", "draw", "show me how it flows", "how does X branch", "graph the", "plot", "over time", "how much of", "gantt", "timeline", "prioritise", "score", "git graph". Write Mermaid or JSON; get aligned text sized for chat or terminal.
---

# Glyphcss diagrams

Author Mermaid or JSON; run the self-contained Node.js 22+ renderer and paste stdout verbatim. Resolve `<skill-dir>` to this file's directory; any working directory works.

```sh
node "<skill-dir>/scripts/render.mjs" - <<'SOURCE'
flowchart LR
  request[Request] --> agent[Agent] --> answer[Answer]
SOURCE
```

## Choose a family

- Dependencies/decisions → flowchart; ordered exchanges → sequence; states of one thing → state diagram.
- Forks/rejoins in histories, plans, stories or lineage → lanes; actual commits → gitGraph; family trees → top-down flowchart.
- Types/members → classDiagram; tables/keys → erDiagram; people/systems → C4; deployed services → architecture; ideas → mindmap; requirements/tests → requirementDiagram; blocks → block.
- Categories → bars; trends → lines/areas; relationships → dots; matrices → cells; proportions → arcs; flows → Sankey; attrition → funnel. Use a table when exact values matter more than shape.
- Dates/dependencies → gantt; dated milestones → timeline; work by status → kanban; judged axes → quadrantChart; criteria → radar; nested shares → treemap; experience → journey; binary fields → packet. Mermaid only.

Read [charts.md](references/charts.md) for JSON charts, per-type rules, sizes, transforms, axes and Mermaid chart syntax. Read [diagrams.md](references/diagrams.md) for JSON graphs/sequences/lanes, Mermaid and layout hints. JSON number arrays, marks, mark arrays or `{marks}` select charts; `nodes`/`edges` select graphs. Mixed graph/chart fields are rejected; author one valid spec. Sequences/lanes need `--kind sequence|lanes`; lanes support `--order oldest-first|newest-first`. Mermaid headers always win over kind/order.

## Target and fit

- Default chat: monochrome 72×24 cells; JSON line/area/bar charts start compact. Never use a tool PTY as the chat viewport or ask for window dimensions.
- Pass a supplied budget with `--width N --height N` (cells, not pixels). Diagrams keep width and may grow taller. `--target terminal` reads stdout's actual TTY minus margins, otherwise 80×24; `--target web` defaults to braille.
- `--charset ascii|box|blocks|braille` overrides defaults. Trends use braille; other chat output uses box. See charts.md for pie/radar/heatmap defaults.
- `--size compact|wide|tall|full` sets 48×16, 72×16, 48×24 or 72×24; explicit dimensions win per axis. Choose JSON chart size from data; if stderr names a larger size, rerun once with it. Mermaid keeps 72×24 unless sized explicitly.
- For `GLYPH_SKILL_UNFIT` diagrams: remove groups/boundaries; then split into overview and 6–8-node details, repeating join nodes; for C4/requirements try a flowchart with the same labelled edges. Shorter labels/flipped direction rarely help.
- For unfit charts: split Sankeys into stage ranges of ≤3 columns with shared nodes repeated; choose a representation supporting every value or split into complete smaller charts. State any split. Preserve values, units, order, series, nodes and connections; never silently aggregate, omit, clip or widen beyond the destination. Disclose a hand-drawn fallback and why.

## Deliver

Stdout is rendered text; stderr is JSON diagnostics. `--json` inspects text/pages, dimensions, budget, metadata and report. Nonzero exit means no usable output. Check labels, axes, units, values and connections; fix source and rerun. Disclose omitted labels/legends, abbreviations, orientation changes or pagination. Paste successful stdout in a fenced `text` block (plain text in terminals), preserving internal spaces and every row/column. Never edit glyphs or substitute source, commands, screenshots or links. Files/other formats only on explicit request; redirect stdout for text exports.

## Report a problem

For persistent reasonable failures or wrong output, offer an issue report using [issues.md](references/issues.md). Recover first; show the draft and file only with explicit approval.

Files in this skill

  • SKILL.md4.9 KB
  • references/charts.md10.9 KB
  • references/diagrams.md11 KB
  • references/issues.md1.9 KB
  • scripts/NOTICE.txt15.5 KB

Attribution

Is this your skill, or is something wrong with this listing? Request removal or report an issue. Author removals are honored within 72 hours.

Comments

Loading comments…

Newsletter

Get the weekly list of safe new skills, and the ones we pulled.