agentleFS
cenconq25cenconq25/claude-code-app-studio/.claude/skills/day-one-patch/SKILL.md

day-one-patch

Skill40 starsChanged 7 months agoMITView on GitHub

Prepare a day-one patch for launch. Scopes, prioritizes, implements, and QA-gates a focused patch addressing issues found between gold master and public launch. Treats the patch as a mini-sprint with its own QA gate and rollback plan.

9 other projects use a version of this.

Adopted by

Most recently taken up by

The file

We split this skill into the rules it gives an agent

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

  • 0

    Common

    Most of the 9 edited versions we read keep these.

  • 2

    Distinctive

    This author's own choices. Worth a look.

  • 4

    Very specific

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

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Use AskUserQuestion to confirm scope:

    line 55
  2. Distinctive:

    Create the patch branch from the release tag:

    line 89
  3. Very specific:

    Run a focused QA cycle scoped to the patch via /team-qa feature: d1-patch.

    line 134
  4. Very specific:

    Document at production/releases/d1-patch-[version]-rollback.md.

    line 159
  5. Very specific:

    Ask before writing to production/releases/d1-patch-[version].md.

    line 205
  6. Very specific:

    If anomalies appear -> /hotfix flow with rollback plan invoked.

    line 254

“Common” is measured: kept by at least 6 in 10 of the 9 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

Day-One Patch

Between the build that went to certification and the build users see at launch, problems show up — late-breaking telemetry, a cert reviewer finding, a marketing copy change. This skill scopes a focused mini- sprint patch, runs it through QA, and lands it on the user before they even notice the gap.


Phase 1: Source the Patch Scope

Resolve the source build:

  • --from=<release-tag> -> read that tag.
  • Otherwise consult production/releases/release-checklist-*.md for the latest release version and use its tag.

Source the candidate patch list from any of:

  • --bug-list=<paths> -> explicit list.
  • Recent BUG records filed since the release tag in production/qa/bugs/.
  • Recent S1/S2 issues triaged in production/qa/triage-*.md.
  • App Store / Play review feedback (if pasted by user).
  • Late-breaking telemetry alerts (Sentry / Crashlytics digest).

Render the candidate list with severity, frequency (if known), and origin source.


Phase 2: Filter to Day-One Scope

Apply the day-one bar — only fixes that meet ALL of these qualify:

  • Severity S1 or S2.
  • Reproducible.
  • Fix is small (< 2 days of engineering, ideally < 1).
  • Fix is low-risk (touches a small, well-understood area).
  • Fix has a regression test plan.

Any candidate failing the bar is either downgraded to a normal sprint, escalated to /hotfix, or rejected.

Use AskUserQuestion to confirm scope:

question: "Day-one scope candidates (top severity first):"
options:
  - "Accept the proposed scope"
  - "Add: [user provides]"
  - "Drop: [user provides]"
  - "Cancel — these are not day-one items"

Phase 3: Risk Assessment

For the agreed scope, run the patch-risk checklist:

  • Each fix has a clear blast radius (which features, which user cohort).
  • Combined diff stays under [project threshold, default 500 LOC].
  • No fix changes the public API surface.
  • No fix changes the on-disk data schema.
  • No fix touches payment flow without a payments specialist on review.
  • No fix touches auth flow without a security review.
  • No fix is bundled with a feature change — patch is fixes only.

Surface any unchecked items and ask the user whether to proceed, descope, or escalate to /hotfix.


Phase 4: Branch and Schedule

Create the patch branch from the release tag:

git checkout -b release/[version]-d1-patch [release-tag]

Ask before running.

Estimate per-bug effort and assemble a patch sprint plan (production/sprints/d1-patch-[version].md):

  • Goals
  • Stories (one per bug)
  • Owner per story
  • Estimated effort
  • QA pass plan
  • Submission deadline (typically the day before launch)

Ask before writing.


Phase 5: Implement Each Story

For each scope item, run the per-story loop:

  1. /create-stories if not already a story.
  2. /story-readiness to validate the story is implementable.
  3. /dev-story to spawn the right specialist and implement.
  4. /code-review on the diff.
  5. /story-done to verify ACs and close.

Each story must add a regression test (this is non-negotiable for a day-one patch — see /regression-suite --add-bug).

Coordinate via Task subagents as needed:

  • lead-developer for general patches.
  • security-engineer for any security-touching fix.
  • qa-tester for test scaffolding.

Phase 6: Patch QA

Run a focused QA cycle scoped to the patch via /team-qa feature: d1-patch.

The cycle includes:

  • Smoke check against the patch branch.
  • Targeted manual walkthroughs for each fix.
  • Regression suite (full).
  • Soak run optional — typically required for memory or background fixes.

Record verdict. Day-one patch demands APPROVED (no conditions). Anything less re-opens the scope.


Phase 7: Rollback Plan

For each fix, capture how to roll back if it misbehaves in production:

  • Revert the patch (delete the patch build, revert to the gold-master build).
  • Per-feature rollback via remote-config kill switch (require these for any non-trivial fix).
  • Pin the underlying flag default to OFF if telemetry shows damage.

Document at production/releases/d1-patch-[version]-rollback.md. Ask before writing.


Phase 8: Compose the Patch Log

# Day-One Patch — [version]

Source build: [release tag]
Patch branch: [branch]
Submitted: [date]
Live: [date]

## Scope
| Story | Bug | Severity | Owner | Risk |

## Diff Summary
- Files changed: [N]
- LOC: [+P / -Q]
- Tests added: [N]

## QA Verdict
- Smoke: [verdict]
- Manual: [PASS/FAIL counts]
- Regression: [verdict]
- Soak: [verdict | not required]
- Sign-off: [verdict]

## Rollback Plan
[summary, link to full doc]

## Approvals
- Engineering Lead
- QA Lead
- Producer
- Release Manager

## Submission
- iOS build: [number]
- Android build: [number]
- Submitted to: [App Store Connect / Play Console]
- Expedited review requested: [yes/no with reason]

Ask before writing to production/releases/d1-patch-[version].md.


Phase 9: Update State

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

## Day-One Patch — [date]
- Version: [version]
- Scope: [N fixes]
- QA verdict: [verdict]
- Rollback: [path]
- Patch log: [path]
- Next: submit, monitor 48h post-launch

Quality Gates / PASS-FAIL

  • PASS — every scope item shipped, regression tests added, QA APPROVED, rollback plan documented, all approvals captured.
  • FAIL — any of: missing regression test, QA short of APPROVED, no rollback for a non-trivial fix, missing approval. Re-scope or escalate.

Examples

Example 1 — Pre-launch d1 patch: 4 candidate bugs reviewed. 2 qualify (S2 paywall copy + S2 push permission timing). Implemented over 2 days, QA APPROVED, rollback flag added. Submitted day before launch with expedited review.

Example 2 — Aborted d1 patch: Initial scope of 7 items violates the < 500 LOC threshold and includes an auth-flow rewrite. Skill flags as risky. User descope: 2 items ship as d1, the auth rewrite goes to next sprint.


Next Steps

  • Submit and request expedited review (App Store) or staged rollout (Play Console).
  • Monitor crash + telemetry for 48 hours post-launch.
  • If anomalies appear -> /hotfix flow with rollback plan invoked.

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.