📊 skillshare v0.18.0 Release Notes

Release date: 2026-03-26
Discussion: #99

TL;DR

skillshare analyze — new command to measure context window token usage per target and lint skill quality
Skill quality lint — 7 built-in rules catch missing fields, short descriptions, and missing trigger phrases
Analyze TUI — interactive exploration with target switching, color dots, sort cycling, and lint icons
Web UI Analyze page — dashboard with chart, table, and token breakdown
Update page polish — sticky search filter and SplitButton actions

Analyze Command

skillshare analyze calculates how much context window budget your skills consume for each configured target. It reports two layers:

Always loaded — the name + description from SKILL.md frontmatter, which AI tools load on every request to decide which skills to trigger
On-demand — the skill body after frontmatter, loaded only when the skill is actually invoked

Token estimates use chars / 4 as an approximation.

skillshare analyze               # interactive TUI (default)
skillshare analyze claude        # details for one target (auto-verbose)
skillshare analyze --verbose     # top 10 largest descriptions per target
skillshare analyze --json        # JSON output for scripting
skillshare analyze -p            # project mode

Passing a target name automatically enables verbose output. Use --no-tui to disable the interactive TUI and print plain text.

Skill Quality Lint

analyze also runs a built-in lint engine against every discovered skill. The engine checks SKILL.md structure and description quality using 7 rules defined in lint_rules.yaml:

Rule	Severity	What it checks
`missing-name`	error	`name` field is empty or missing
`missing-description`	error	`description` field is empty or missing
`empty-body`	error	Skill body (after frontmatter) is empty
`description-too-short`	warning	Description under 50 characters
`description-too-long`	warning	Description exceeds 1024-character limit
`description-near-limit`	warning	Description between 900–1024 characters
`no-trigger-phrase`	warning	Description lacks trigger phrases like "Use when…"

Lint issues appear in the TUI next to skill names (✗ for errors, ⚠ for warnings) and in --json output as lint_issues per skill. This makes it easy to find and fix skills that AI tools might not trigger correctly.

Interactive TUI

The default analyze mode launches a full-screen bubbletea TUI with:

Left panel — skill list sorted by token cost, with color-coded dots (red = P75+, yellow = P25–P75, green = below P25)
Right panel — detail view showing token breakdown, lint quality issues, file path, tracked status, and description preview
Bottom bar — target selector and token totals

Targets with identical skill sets are merged into groups (e.g., "claude, cursor" shown as one tab). Press Tab/Shift+Tab to switch, / to filter, s to cycle sort order (tokens↓ → tokens↑ → name A→Z → Z→A), and Ctrl+d/Ctrl+u to scroll the detail panel.

Web UI Analyze Page

The web dashboard adds a new Analyze page accessible from the sidebar navigation. It shows:

Per-target token usage summary with a chart
Skill table with description/body character and token counts
Lint issue indicators per skill
Token breakdown on individual Skill Detail pages

A new GET /api/analyze endpoint powers the page, returning per-target context analysis including lint issues, skill paths, tracked status, and descriptions.

Update Page Improvements

The Update page gets two usability improvements:

Sticky search filter — the search input sticks to the top when scrolling long skill lists
SplitButton actions — action buttons are now a compact SplitButton component instead of separate buttons

Unified Dialog Styling

All modal dialogs across the Web UI now share consistent styling via a DialogShell component, replacing per-dialog styling that had minor visual inconsistencies.

Changelog

a636fc6 feat(analyze): add --no-tui flag and TUI dispatch
10415a7 feat(analyze): add Quality section and lint icons to TUI
439abbd feat(analyze): add TUI list delegate with color dots and bar chart
956a719 feat(analyze): add TUI-only fields to analyzeSkillEntry
620141a feat(analyze): add context analysis command
2e43e37 feat(analyze): add lint rules YAML and LintSkill engine
64ed9b0 feat(analyze): add malformed-frontmatter lint rule and polish UI/TUI
e88c32b feat(analyze): implement interactive TUI with target switching and detail panel
2811a19 feat(analyze): return description from calcContextFromContent for TUI
57ce3f5 feat(analyze): wire lint into discovery and analyze data flow
fe3e72f feat(ui): add GET /api/analyze endpoint
920feab feat(ui): add analyze API types, route, and nav item
977d761 feat(ui): add sticky search filter to Update page
ae39892 feat(ui): add token breakdown to skill detail and polish AnalyzePage
9f38326 feat(ui): implement full AnalyzePage with dashboard, chart, and table
9ecd78e feat(ui): redesign AnalyzePage and unify dialog styling across all modals
89df6d2 fix(ui): center pin decoration on playful theme detail cards
bb30924 refactor(analyze): merge identical targets into groups in TUI
78ba5c7 refactor(analyze): polish TUI rendering and add docs
deb834c refactor(analyze): remove Context category bar from TUI
d377e6b refactor(analyze): simplify lint engine and fix review findings
d58b272 refactor(ui): replace Update page action buttons with SplitButton
68111d2 refactor: rename project skills with skillshare- prefix
d404eda test(analyze): add integration test for JSON lint output
42dd4e4 test(analyze): add sort cycling and target switch unit tests
f5e5588 test(analyze): add unit tests for all 7 lint rules

runkids/skillshare v0.18.0 on GitHub