check-ghostty-update
Skill787 starsChanged 3 months agoMITView on GitHub
Check whether ghostty (the terminal core termio embeds via the termio-sh/libghostty-swift fork) has shipped anything worth pulling since the binary termio is currently pinned to. Resolves the pinned ghostty sha, diffs it against ghostty main, filters the gap for changes that actually reach termio's embedded VT/terminal core (dropping app-chrome / packaging / i18n noise), and gives a keep-waiting-or-pull-now verdict. Invoke when the user says 'check ghostty update', 'is ghostty out of date', 'did ghostty ship anything major', 'should I bump ghostty/libghostty', '看看 ghostty 有没有更新', '检查 ghostty 更新', 'ghostty 有什么重大更新', or '要不要更新 libghostty'.
In this repository29 other files it gives its agents.
- AGENTS.mdAGENTS.md
- CLAUDE.mdCLAUDE.md
- animation-vocabularySkill
- apple-designSkill
- app-screenshot-debugSkill
- ascSkill
- bump-versionSkill
- code-cleanSkill
- conventional-commitSkill
- dia-source-analysisSkill
- docSkill
- emil-design-engSkill
- find-animation-opportunitiesSkill
- improve-animationsSkill
- ios-app-analysisSkill
- ios-rebuild-devSkill
- issue-creatorSkill
- macos-rebuild-devSkill
- native-app-profilingSkill
- og-generationSkill
- pick-ui-librarySkill
- prototypeSkill
- review-animationsSkill
- review-copySkill
- Swift Performance Optimization SkillSkill
- performance-auditSkill
- termio-bug-reportSkill
- termioSkill
- testflight-releaseSkill
The file
We split this skill into the rules it gives an agent
3 rules. About 0 are common, 0 are distinctive, and 3 are very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈0
Distinctive · estimate
This author's own choices. Worth a look.
3
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Very specific: line 25
Run these steps and report.
- Very specific: line 47
If any of these come back empty, fall back: the fork's latest storage.* release is usually what a fresh resolve would pick — gh release …
- Very specific: line 85
Read the survivors yourself — the grep is a first pass, not the judgement.
≈ “Common” is an estimate: at least 1 in 5 of the 350,495 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
Check ghostty update
termio does not build ghostty itself — it embeds ghostty's terminal core through
the termio-sh/libghostty-swift fork, whose weekly CI rebuilds GhosttyKit.xcframework
against ghostty main and publishes it as a storage.X.Y.Z release ([[termio-libghostty-swift]]).
So "is ghostty up to date" really means: how far is ghostty main ahead of the sha baked
into the binary termio is pinned to, and is anything in that gap relevant to termio?
termio is an embedder, so only a slice of ghostty's changes matter. Weight the diff accordingly — see Signal vs. noise below. Local self-build is impossible on this Mac (Tahoe/zig deadlock); updating the binary always goes through the fork's CI.
What this skill does NOT do
Report only. It does not tag the fork, trigger CI, or bump any pin. Those are the follow-up the verdict points to (and separate skills / explicit asks).
Procedure
Run these steps and report; everything is read-only gh + grep.
1. Resolve the ghostty sha termio is currently on
The pin chain is: termio Package.resolved → fork version tag → that tag's Package.swift
names a storage.X.Y.Z binary → that storage release's title carries the ghostty sha
(format ... · ghostty v1.3.1-<N>-g<sha>).
REPO=$(git rev-parse --show-toplevel)
FORK=termio-sh/libghostty-swift
GHOSTTY=ghostty-org/ghostty
VER=$(grep -A6 '"identity" : "libghostty-swift"' "$REPO/Package.resolved" \
| grep '"version"' | head -1 | sed -E 's/.*"version" : "([^"]+)".*/\1/')
STORAGE=$(gh api "repos/$FORK/contents/Package.swift?ref=$VER" --jq '.content' \
| base64 -d | grep -oE 'storage\.[0-9]+\.[0-9]+\.[0-9]+' | head -1)
NAME=$(gh release view "$STORAGE" --repo "$FORK" --json name --jq '.name')
CUR=$(echo "$NAME" | grep -oE 'g[0-9a-f]{7,}' | sed 's/^g//')
echo "termio pin: fork $VER → $STORAGE → ghostty $NAME (sha $CUR)"If any of these come back empty, fall back: the fork's latest storage.* release is
usually what a fresh resolve would pick —
gh release list --repo $FORK | grep storage | head -1 — and note the discrepancy.
2. Diff against ghostty main
gh api "repos/$GHOSTTY/compare/$CUR...main" \
--jq '"ahead: \(.ahead_by), head: \(.commits[-1].sha[0:7]) @ \(.commits[-1].commit.author.date[0:10])"'ahead: 0→ you are current. Stop and report "nothing to pull."- Otherwise capture
ahead_byand the head sha/date for the report.
3. Pull the gap and separate signal from noise
gh api "repos/$GHOSTTY/compare/$CUR...main" \
--jq '.commits[].commit.message | split("\n")[0]' > /tmp/ghostty_gap.txt
wc -l /tmp/ghostty_gap.txtSignal — changes that reach termio's embedded core (surface these):
prefixes lib-vt:, terminal:, terminal/…:, renderer:, font:, kitty (graphics /
keyboard / protocol), osc, selection, sixel, hyperlink, unicode, perf /
throughput, scrollback, resize/reflow, mode/DEC private modes.
Noise — termio host-manages IO and ships its own AppKit/SwiftUI chrome, so ignore:
macOS: (splits, tabs, title bar, NSScrollPocket, HIG), gtk/gnome, pkg/… /
apple-sdk / build.zig, i18n/translation, nushell/shell-integration, docs,
ci, gitignore, VOUCHED/changelog.
grep -iE 'lib-vt|terminal[:/]|renderer|kitty|osc|selection|sixel|hyperlink|unicode|scrollback|reflow|throughput|perf|font' /tmp/ghostty_gap.txt \
| grep -ivE 'macos:|gtk|gnome|pkg/|apple-sdk|i18n|translat|nushell|docs:|^ci|gitignore|vouched' \
| sort -uRead the survivors yourself — the grep is a first pass, not the judgement. For any that
look big (memory/perf wins, a new VT/embedder API termio could use, a protocol termio
renders like kitty graphics), confirm it isn't already in the pinned binary by date
(a commit older than the binary's build date is already in) or with a compare:
gh api repos/$GHOSTTY/compare/<that-sha>...$CUR --jq .status → ahead/identical = already have it.
4. Verdict
Give a short, honest call:
- Nothing relevant (gap is all app-chrome/packaging/i18n) → "no reason to bump; the Monday cron will roll it forward on its own." This is the common case — ghostty is high-velocity but most of it is app-layer.
- Something relevant but minor → name it, say it'll arrive with the next weekly build, no rush.
- Something genuinely major (big perf/memory win, a VT/embedder API termio wants, a correctness fix in a path termio exercises) → name it with its PR #, and offer the pull-now path.
Pull-now path (only when the verdict warrants it)
The binary can't be built locally. To pull ghostty early:
- Manually dispatch the fork's CI against a specific ghostty ref — the fork's
.github/workflows/build.ymltakes aghostty_refinput (default buildsmainHEAD):gh workflow run build.yml --repo termio-sh/libghostty-swift -f ghostty_ref=main(or pin an exact sha). It publishes a newstorage.X.Y.Z+X.Y.Ztag (~30 min). - Bump termio's pin (
Package.swiftfrom:+ re-resolvePackage.resolved), and the iOSTermioMobile.xcodeprojpin if the phone should get it too (it pins the fork separately — see [[termio-libghostty-swift]]). macos-rebuild-devand verify.
Notes
- ghostty cuts stable tags rarely (last was
v1.3.1, 2026-03), and the fork tracksmain, so termio typically rides ~1400+ commits ahead of ghostty's newest stable tag. "No new ghostty release" ≠ "nothing new" — always compare againstmain, never against tags. - The fork's own semver (
1.0.x) is the wrapper version and is decoupled from ghostty's version; a wrapper-only change (like the Swift-sideInMemoryTerminalSessionfixes) bumps the fork tag while reusing the samestorage.*ghostty binary. That's why step 1 reads the storage tag from the pinned tag'sPackage.swiftrather than assumingstorage.<same-version>. - Needs the
ghCLI authenticated. All calls are read-only.
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
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.

