agentleFS
egregore-labsegregore-labs/egregore/.claude/skills/tui-design/SKILL.md

tui-design

Skill293 starsChanged 11 days agoMITView on GitHub

TUI Design System for Egregore slash commands. Use when designing or modifying any command's terminal output to ensure visual consistency. Triggers on: TUI, terminal output, box drawing, command output formatting, sigils, section markers. Do NOT use for browser rendering (use dark-mode) or external prose (use egregore-voice).

12 other projects use a version of this.

Adopted by

Most recently taken up by

Used with

Projects with it also use unslop.

Other files people use alongside this one.

Files found in the same public projects as a version of this one, more often than they turn up anywhere else. Each project counts once, large collections count for less, and this file's own repository is left out.
  • photon-groveSkill

    unslop

    Detects and rewrites generic, overly polished, or AI-sounding prose while preserving meaning. Use when the user asks to "unslop", humanize, de-AI, tighten, or review whether writing sounds generated.

    Also in 3 projects that use this

The file

We split this skill into the rules it gives an agent

10 rules. 0 are common, 3 are distinctive, and 7 are very specific to this project.

  • 0

    Common

    Most of the 12 edited versions we read keep these.

  • 3

    Distinctive

    This author's own choices. Worth a look.

  • 7

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Very specific:

    Not this: browser rendering → dark-mode · external prose → egregore-voice

    line 19
  2. Very specific:

    Read org name from jq -r '.org_name' egregore.json.

    line 29
  3. Very specific:

    Header separator: always use ├───┤ between header row and content.

    line 75
  4. Very specific:

    Section dividers: use ├───┤ between logical groups.

    line 77
  5. Distinctive:

    All commands use full-width, single-column, stacked sections.

    line 81
  6. Very specific:

    All content lines must have matching left and right │ borders

    line 86
  7. Very specific:

    When a command's TUI includes interactive follow-up (like /activity numbered items):

    line 174
  8. Very specific:

    Add footer text inside the box: Type a number to act, or keep working.

    line 177
  9. Distinctive:

    Ask with a structured question when available and permitted in this session.

    line 178
  10. Distinctive:

    No bold, italic, or color codes in the markdown instruction files

    line 188

“Common” is measured: kept by at least 6 in 10 of the 12 edited versions of this file we could read. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

TUI Design System

Voice Complex

This skill governs TERMINAL RENDERING — structural, not tonal. Shared aesthetic: voice-bedrock (.claude/rules/voice-bedrock.md). For product voice in command output prose: product-voice skill. For browser visual output: dark-mode skill, which takes every color from the design system registry (packages/design-system/themes/registry.json). For ASCII diagrams in explanations: visual-explain skill.

When to invoke

User says: "TUI", "terminal output", "box drawing", "command output formatting", "sigils", "section markers", or is designing/modifying any slash command's terminal output. Not this: browser rendering → dark-mode · external prose → egregore-voice

1. Brand Elements

Command Headers

/activity is the flagship — it gets the full branded header with org name:

{ORG_NAME} EGREGORE ✦ ACTIVITY DASHBOARD

Read org name from jq -r '.org_name' egregore.json. The ✦ star is reserved exclusively for the activity dashboard and Egregore-wide branding.

Other commands get sigil + command name only — no org name:

CommandHeaderSigil Meaning
/activity{ORG_NAME} EGREGORE ✦ ACTIVITY DASHBOARDThe star — awareness, the full field
/ask? ASKThe question — direct, clear inquiry
/reflect◎ REFLECTIONThe eye — looking inward, seeing clearly
/deep-reflect◈ DEEP REFLECTThe faceted eye — looking inward with structure
/handoff⇌ HANDOFFThe bridge — passing context between minds
/quest⚑ QUESTThe flag — planting direction in unknown territory
/todo□ TODOThe checkbox — actionable items, personal intent
/archive◇ ARCHIVEThe diamond — crystallized pattern, refined and preserved
/eval⊕ EVALThe crosshair — measuring, comparing, calibrating

Section Markers

Used within all TUI boxes to prefix items by category:

  • ● Action items / things needing attention
  • ✓ Completed / received / done items
  • → Navigation pointers (what to do next, entry points)
  • ◦ Informational items (sessions, quests, neutral data)
  • ⚑ Quests
  • ⇌ Handoffs (in activity action items)
  • ◎ Reflections
  • □ Todos (in activity dashboard)
  • ★ Priority marker (todo items with priority >= 2)
  • ◉ Artifacts (in handoff summaries)

2. Box Drawing Rules

Line Patterns (CRITICAL)

LLMs cannot count character widths. Do NOT draw boxes freehand. Instead, produce exactly 4 line patterns. All commands use 72-char outer width (N = 70 fill characters).

  1. Top: ┌ + 70×─ + ┐
  2. Separator: ├ + 70×─ + ┤
  3. Content: │ + 2 spaces + text + pad to 68 text chars + │
  4. Bottom: └ + 70×─ + ┘

Characters: ┌ top-left, ─ horizontal, ┐ top-right, │ vertical, ├ left-tee, ┤ right-tee, └ bottom-left, ┘ bottom-right.

Separator lines are always identical — copy-paste the same 72-char string every time. Content lines have ONLY the outer │ as borders. Pad every content line with trailing spaces so the closing │ lands at position 72.

Header separator: always use ├───┤ between header row and content.

Section dividers: use ├───┤ between logical groups. No sub-boxes — never use ┌─┐/└─┘ inside the outer frame. LLMs cannot reliably align nested borders.

Layout

All commands use full-width, single-column, stacked sections. No column splits — they cause truncation and render poorly. Content flows top-to-bottom with section headers separating groups.

3. Width Standards

  • All commands: ~72 characters outer width
  • All content lines must have matching left and right │ borders

4. Content Formatting

Session Lines (Compact)

Feb 07  Topic description

Date is always Mon DD format, right-padded to 6 chars. Topic follows after 2 spaces.

Team Session Lines

Feb 07  oz: Topic description

Person name prefixed with colon separator.

Numbered Action Items

[1] Description of actionable thing
[2] ⇌ Handoff from ali: blog styling (yesterday)

Numbers in brackets. Use section markers after the bracket for typed items.

Status Confirmations

✓ Saved to knowledge/decisions/2026-02-08-...
✓ Memory index updated
✓ Auto-saved

Progress Steps

[1/5] ✓ Conversation file
[2/5] ✓ Index updated
[3/5] ✓ Session saved to memory
[4/5] ✓ Pushed + PR created
[5/5] ✓ Oz notified

Linked Items

◦ Quest: research-agent
◦ Project: egregore

Entry Points

→ memory/knowledge/decisions/2026-02-07-defensibility-...
→ memory/knowledge/decisions/2026-02-07-business-model.md

Truncation

  • Topics longer than 35 chars in column layouts get ... suffix
  • File paths: shorten to last 2 segments when possible
  • Never break mid-word

5. Section Visibility Rules

  • Empty sections are omitted entirely — no placeholder text, no "none yet", no empty boxes
  • Single action item still gets the action items box — consistency over minimalism
  • Section headers only appear when their section has content

6. Edge Cases

ScenarioHandling
No sessions yetShow Projects section instead (first-timer experience)
Neo4j downFile-based fallback, simpler layout, no columns
Single action itemStill show action items box
No team activityOmit TEAM section
Very long org nameTruncate at 20 chars in activity header
No questsOmit quests section
No PRsOmit PRs section
No action items + no answersOmit both sections, go straight to columns

7. Anti-Patterns

  • Avoid sub-boxes entirely — use ├───┤ flat dividers instead (LLMs can't count widths for nested borders)
  • Never use tables with │ column separator AND sub-boxes in the same section (pick one)
  • Never show empty section headers — omit the entire section
  • Never use raw file paths in TUI output — use → pointer with shortened path
  • Never break box characters across lines — each line must have valid left/right borders
  • Never use the ✦ star outside /activity header
  • Never include org name in non-activity command headers
  • Never use emoji in TUI boxes (sigils and markers only)

8. Structured questions

When a command's TUI includes interactive follow-up (like /activity numbered items):

  1. Display the full TUI box first
  2. Add footer text inside the box: Type a number to act, or keep working.
  3. Ask with a structured question when available and permitted in this session; otherwise ask in plain text. Use options matching the numbered items
  4. Include a "Skip" option for non-action

9. Color and Emphasis

Terminal output is monospace plain text. Emphasis is structural only:

  • Section headers: ALL CAPS
  • Sigils and markers provide visual weight
  • Sub-boxes create grouping
  • Indentation creates hierarchy
  • No bold, italic, or color codes in the markdown instruction files

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.