agentleFS
cenconq25cenconq25/claude-code-app-studio/.claude/skills/hotfix/SKILL.md

hotfix

Skill40 starsChanged 7 months agoMITView on GitHub

Emergency fix workflow that bypasses normal sprint process while keeping a full audit trail. Creates the hotfix branch, tracks approvals, ensures backport to main, and produces release-store-ready notes. Use only for production-impacting S1 bugs or critical store-rejection issues.

29 other projects use a version of this.

Adopted by

Most recently taken up by

Used with

Projects with it also use playtest-report, team-audio and 2 more.

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.
  • CoralGameSkill

    playtest-report

    Generates a structured playtest report template or analyzes existing playtest notes into a structured format. Use this to standardize playtest feedback collection and analysis.

    Also in 12 projects that use this

  • CoralGameSkill

    team-audio

    Orchestrate audio team: audio-director + sound-designer + technical-artist + gameplay-programmer for full audio pipeline from direction to implementation.

    Also in 12 projects that use this

  • +2 more from CoralGame

The file

Show its rules and text

We split this skill into the rules it gives an agent

16 rules. 0 are common, 10 are distinctive, and 6 are very specific to this project.

  • 0

    Common

    Most of the 27 edited versions we read keep these.

  • 10

    Distinctive

    This author's own choices. Worth a look.

  • 6

    Very specific

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

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Ask the user via AskUserQuestion to confirm the trigger:

    line 28
  2. Very specific:

    If [D], stop and recommend /bug-report or /bug-triage.

    line 35
  3. Very specific:

    If BUG-ID is given, read production/qa/bugs/BUG-[ID]-*.md.

    line 41
  4. Distinctive:

    If any approval is missing, stop.

    line 63
  5. Distinctive:

    Create the hotfix branch from that tag:

    line 82
  6. Very specific:

    Ask before running. If the user prefers a different base (a release branch, e.g., release/2.5.x), use that.

    line 88
  7. Distinctive:

    Create the minimal change.

    line 112
  8. Very specific:

    Add a regression test (this is non-negotiable for hotfix — see /regression-suite --add-bug).

    line 113
  9. Distinctive:

    Document why the fix works in the bug's investigation log.

    line 115
  10. Very specific:

    Run the smoke check against the hotfix branch via /smoke-check.

    line 121
  11. Distinctive:

    Run the regression suite via the project's test runner:

    line 123
  12. Distinctive:

    If anything fails, stop. The hotfix must not introduce new red.

    line 132
  13. Distinctive:

    If main has diverged so much that the cherry-pick is non-trivial, surface that and consider a separate "main-equivalent fix" PR.

    line 151
  14. Distinctive:

    Ask the user to confirm the PR destinations and reviewers before opening PRs.

    line 154
  15. Distinctive:

    Update the bug record:

    line 225
  16. Very specific:

    If invoked with --rollback, the hotfix is being reversed.

    line 235

“Common” is measured: kept by at least 6 in 10 of the 27 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

Hotfix

The break-glass procedure for shipping a fix outside a sprint. Trades process for speed but keeps an audit trail tight enough to survive post-incident review.


Phase 1: Confirm This Is Actually a Hotfix

Hotfix is justified only when ALL of these hold:

  • A production build is currently live and impacted.
  • The issue is S1 OR a store-rejection blocker OR a security exposure.
  • Waiting for the next sprint would cause further user harm.

Otherwise, route to /bug-triage for normal scheduling.

Ask the user via AskUserQuestion to confirm the trigger:

  • [A] Live S1 — users impacted right now
  • [B] App Store / Play Store rejection — review board needs a fix
  • [C] Security disclosure — needs immediate patch
  • [D] Cancel — this is not a hotfix

If [D], stop and recommend /bug-report or /bug-triage.


Phase 2: Resolve the Bug Record

If BUG-ID is given, read production/qa/bugs/BUG-[ID]-*.md. If not, prompt the user to file one via /bug-report first — every hotfix must have a tracked bug.

Read:

  • The bug record in full.
  • Any related ADRs from the bug's "Related" section.
  • The most recently shipped build's tag (via git tag --sort=-creatordate).

Phase 3: Approval Gate

Hotfix bypasses sprint planning, so a higher approval bar applies. Use AskUserQuestion to capture each:

  • Engineering Lead approval (name)
  • Producer approval (name)
  • QA Lead approval (name)
  • Release Manager approval (name)

If any approval is missing, stop. The audit trail requires all four for a true hotfix. (The user can choose to proceed without a formal hotfix process — but then it's a normal sprint fix.)

Record approvals into the hotfix log (created in Phase 8).


Phase 4: Branch Strategy

Determine the source. Use Bash:

git fetch --tags
git tag --sort=-creatordate | head -n 5

Confirm with the user which release tag is currently live in production.

Create the hotfix branch from that tag:

git checkout -b hotfix/BUG-[ID]-[short-slug] [release-tag]

Ask before running. If the user prefers a different base (a release branch, e.g., release/2.5.x), use that.


Phase 5: Implement the Fix

Spawn the right specialist via Task. Routing matches /dev-story:

  • Crash in core Logic -> lead-developer + framework specialist.
  • UI regression -> lead-developer + framework UI specialist.
  • Backend / sync / data -> backend-engineer.
  • Security exposure -> security-engineer.

Pass:

  • The bug record.
  • The reproduction steps.
  • The release tag the fix is based on.
  • Explicit constraint: "Smallest possible diff. No refactors. No tangential improvements. The fix must be backportable to main."

The agent should:

  • Create the minimal change.
  • Add a regression test (this is non-negotiable for hotfix — see /regression-suite --add-bug).
  • Document why the fix works in the bug's investigation log.

Phase 6: Hotfix Verification

Run the smoke check against the hotfix branch via /smoke-check.

Run the regression suite via the project's test runner:

  • npx jest tests/regression
  • flutter test test/regression
  • xcodebuild test -only-testing:RegressionTests
  • ./gradlew testDebugUnitTest --tests "*regression*"

Capture per-test results.

If anything fails, stop. The hotfix must not introduce new red.

For visual changes, also walk through the affected flow on at least one Tier A device per platform with the user.


Phase 7: Backport Plan

Hotfixes must land on both the release branch (or tag) AND main.

Capture the plan:

  1. Merge the hotfix branch into the release branch (or create the release branch from the tag if none exists).
  2. Cherry-pick the hotfix commit(s) onto main.
  3. Resolve conflicts on main if main has diverged.
  4. Open PRs for both targets.

If main has diverged so much that the cherry-pick is non-trivial, surface that and consider a separate "main-equivalent fix" PR. Both must land.

Ask the user to confirm the PR destinations and reviewers before opening PRs. (This skill does not push to remote — leave that to the user.)


Phase 8: Compose the Hotfix Log

Write production/releases/hotfixes/HOTFIX-[date]-BUG-[ID].md:

# Hotfix Log: BUG-[ID] — [Title]

**Date**: [date]
**Trigger**: [S1 / store rejection / security]
**Affected build**: [version]
**Hotfix build target**: [version]

## Approvals
- Engineering Lead: [name]
- Producer: [name]
- QA Lead: [name]
- Release Manager: [name]

## Timeline
- [time] Bug filed
- [time] Hotfix declared
- [time] Branch created from [tag]
- [time] Fix implemented in [commit]
- [time] Regression test added at [path]
- [time] Smoke check PASS
- [time] Regression suite PASS
- [time] Manual verification on [devices]
- [time] PR opened: [release branch URL]
- [time] PR opened: [main URL]

## Diff
- Files changed: [count]
- Lines: [+N / -M]
- Test added: [path]

## Risk Assessment
- Blast radius: [scope]
- Rollback path: [how to roll back]
- Expected user impact post-deploy: [statement]

## Rollback Plan
[explicit steps if the hotfix itself misbehaves]

## Post-Mortem Action
[link to scheduled retrospective or "Not required — minor scope"]

Ask before writing.


Phase 9: Update State

Append to production/session-state/active.md:

## Hotfix — [date]
- Bug: BUG-[ID]
- Branch: hotfix/...
- Smoke: PASS
- Regression: PASS
- PRs open: [release], [main]
- Hotfix log: [path]
- Next: merge release PR, deploy expedited build, then merge main PR

Update the bug record:

  • Status: In Hotfix while in flight.
  • Status: Closed once both PRs are merged AND the hotfix build is live.

Rollback Mode

If invoked with --rollback, the hotfix is being reversed. Workflow:

  1. Confirm the hotfix tag is live and is causing harm.
  2. Open a revert PR on the release branch.
  3. Re-run smoke + regression on the reverted state.
  4. Append a ## Rollback block to the hotfix log capturing why.
  5. Re-open the underlying bug as Reopened and re-triage via /bug-triage.

Quality Gates / PASS-FAIL

  • PASS — fix landed, regression test added, smoke + regression suite green, both PRs open, log written, all four approvals captured.
  • FAIL — any of: missing approval, no regression test, smoke or regression red, log incomplete. Do not declare the hotfix shipped.

Examples

Example 1 — Live S1 crash on Android resume: Bug filed, approvals captured, branch off v2.5.0, fix is a 4-line Compose change, regression test added, smoke + regression PASS, PRs opened to release/2.5.x and main. Hotfix log written.

Example 2 — Store rejection for ATT prompt timing: Apple rejects build for showing tracking dialog before main UI loads. Hotfix delays the prompt to first relevant moment. Test added. Smoke PASS. New build submitted with the hotfix log attached to the appeal.


Next Steps

  • Merge release PR, ship the hotfix build via /team-release (in expedited mode).
  • Merge main PR.
  • Schedule a post-mortem if the hotfix involved a regression caused by recent work — /retrospective after the next sprint.

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.