agentleFS
LedgerHQLedgerHQ/ledger-live/.agents/skills/run-builds/SKILL.md

run-builds

Skill627 starsChanged 10 days agoMITView on GitHub

Trigger the on-demand production build workflows on LedgerHQ/ledger-live-build (Desktop, Android APK, iOS) for a ledger-live branch, then post a PR comment. Use when asked to "run builds", "produce a build", "build the app(s)", or "make an APK/iOS/desktop build" for a PR/branch.

The file

We split this skill into the rules it gives an agent

4 rules. About 0 are common, 2 are distinctive, and 2 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈2

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 2

    Very specific

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

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    If the user explicitly names a platform ("make an APK", "iOS build"), honor that and skip the inference.

    line 28
  2. Very specific:

    All three share repo (default LedgerHQ/ledger-live) and ref (default develop).

    line 51
  3. Very specific:

    If the branch has an open ledger-live PR, post the dispatched runs on it.

    line 77
  4. Distinctive:

    Only include rows for the platforms you actually dispatched.

    line 100

≈ “Common” is an estimate: at least 1 in 5 of the 364,980 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.

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

Produce app builds on ledger-live-build CI

When: to produce installable artifacts (desktop installers, APK, iOS) of a ledger-live branch — for QA, manual testing, or sharing a build with reviewers.

These workflows live in LedgerHQ/ledger-live-build, not ledger-live. That changes how you dispatch them vs. run-e2e-ci:

  • Target the build repo: gh workflow run <wf> -R LedgerHQ/ledger-live-build.
  • The branch you want to build is passed as inputs, not --ref: -f repo=LedgerHQ/ledger-live -f ref=<branch>. The workflow checks out inputs.ref from inputs.repo. --ref (if given) would select the ledger-live-build ref and should be left at default (main).
  • gh must run outside the sandbox (dangerouslyDisableSandbox: true).

1. Scope by context — only build what the diff impacts

Builds are slow and expensive (especially desktop = macOS + Linux + Windows). Study the PR diff / branch changes and dispatch only the relevant surfaces:

Diff touchesBuild
apps/ledger-live-desktop/**Desktop
apps/ledger-live-mobile/**Android + iOS
a package under shared/, domain/, features/, devtools/the app(s) that depend on it, see below
support/** only (development-only presets)nothing
shared libs (libs/**, e.g. ledger-live-common, coin-modules) impacting both appsDesktop + Android + iOS

If the user explicitly names a platform ("make an APK", "iOS build"), honor that and skip the inference. When in doubt about a shared-lib change, ask which app(s) to build rather than building all three.

Derive the app from a change in the layers with the Nx graph. Pass the changed source files; a package depended on by one app only reaches that app (@features/flow-app-unlock → mobile, @features/platform-app-update → desktop):

FILES=$(gh pr view <pr> -R LedgerHQ/ledger-live --json files \
  --jq '[.files[].path | select(test("^(shared|domain|features|devtools)/") and (endswith("package.json") | not))] | join(",")')
pnpm nx show projects --affected --files="$FILES" --json \
  | jq -r '.[] | select(. == "ledger-live-desktop" or . == "live-mobile")'

ledger-live-desktop → Desktop, live-mobile → Android + iOS. Then narrow by platform suffix, which Nx does not see: a change only in *.web.ts(x) files ships on desktop, only in *.native.ts(x) on mobile. A package.json change marks every dependent app as affected; read what it changes before building all three. A devtools/ change only shows in the developer tools screen; build it only when someone needs to test that screen.

Scope off the PR's file list, not git diff develop...HEAD. A stale merge-base makes the local three-dot diff pull in hundreds of unrelated files. When a PR exists, use gh pr view <pr> -R LedgerHQ/ledger-live --json files --jq '.files[].path' for the true change set.

2. Workflows & inputs

WorkflowFileProducesExtra inputs
Desktopbuild-desktop.ymlmacOS, Linux, Windows bundlesupload_datadog_sourcemaps (bool, default false)
Androidbuild-apk.yml.apk—
iOSbuild-ios.ymliOS build—

All three share repo (default LedgerHQ/ledger-live) and ref (default develop).

3. Dispatch

# Branch to build: prefer the PR head (robust to detached HEAD on a PR SHA),
# fall back to the current branch only when there's no PR in context.
BR=$(gh pr view <pr> -R LedgerHQ/ledger-live --json headRefName --jq .headRefName 2>/dev/null \
     || git rev-parse --abbrev-ref HEAD)
R=LedgerHQ/ledger-live-build

gh workflow run build-desktop.yml -R "$R" -f repo=LedgerHQ/ledger-live -f ref="$BR"
gh workflow run build-apk.yml     -R "$R" -f repo=LedgerHQ/ledger-live -f ref="$BR"
gh workflow run build-ios.yml     -R "$R" -f repo=LedgerHQ/ledger-live -f ref="$BR"

Then grab the run IDs (dispatch is async; give it a few seconds). Filter per workflow since run-names don't echo the source branch:

gh run list -R "$R" --workflow build-desktop.yml --limit 3
gh run list -R "$R" --workflow build-apk.yml     --limit 3
gh run list -R "$R" --workflow build-ios.yml     --limit 3

4. Comment on the PR (if a PR is in context)

If the branch has an open ledger-live PR, post the dispatched runs on it. The PR lives in ledger-live, the runs in ledger-live-build — link accordingly. Use --edit-last to update an existing comment instead of stacking.

# First time: post a new comment.
gh pr comment <pr> -R LedgerHQ/ledger-live --body "..."
# On re-dispatch: update the same comment instead of stacking a new one.
gh pr comment <pr> -R LedgerHQ/ledger-live --edit-last --body "..."
## 📦 Triggered builds

Dispatched on `ledger-live-build` for `<branch>` (commit <short-sha>):

| Platform | Run |
|---|---|
| Desktop (macOS / Linux / Windows) | [run <id>](https://github.com/LedgerHQ/ledger-live-build/actions/runs/<id>) |
| Android (APK) | [run <id>](https://github.com/LedgerHQ/ledger-live-build/actions/runs/<id>) |
| iOS | [run <id>](https://github.com/LedgerHQ/ledger-live-build/actions/runs/<id>) |

Artifacts will be attached to each run on completion.

Only include rows for the platforms you actually dispatched. If there's no PR in context, just report the run links back in chat.

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.