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

gate-check

Skill40 starsChanged 7 months agoMITView on GitHub

Validate readiness to advance between development phases. Synthesizes QA, perf, security, asset, balance, and localization verdicts into a single PASS/CONCERNS/FAIL with specific blockers and required artifacts. Use when asking 'are we ready to advance?'.

20 other projects use a version of this.

Adopted by

Most recently taken up by

Used with

Projects with it also use setup-engine, team-level and 1 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

    setup-engine

    Configure the project's game engine and version. Pins the engine in CLAUDE.md, detects knowledge gaps, and populates engine reference docs via WebSearch when the version is beyond the LLM's training data.

    Also in 13 projects that use this

  • CoralGameSkill

    team-level

    Orchestrate level design team: level-designer + narrative-director + world-builder + art-director + systems-designer + qa-tester for complete area/level creation.

    Also in 12 projects that use this

  • +1 more from CoralGame

The file

We split this skill into the rules it gives an agent

7 rules. 0 are common, 4 are distinctive, and 3 are very specific to this project.

  • 0

    Common

    Most of the 19 edited versions we read keep these.

  • 4

    Distinctive

    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
  1. Distinctive:

    No argument — ask the user.

    line 28
  2. Very specific:

    All MVP PRDs reviewed via /review-all-prds with PASS verdict.

    line 67
  3. Distinctive:

    All Critical / Must-Have features implemented per PRD.

    line 82
  4. Distinctive:

    Look for contradictions between artifacts:

    line 114
  5. Distinctive:

    All required-for-this-gate artifacts at +1, no contradictions -> PASS.

    line 139
  6. Very specific:

    Ask before writing to production/gates/gate-[from-to]-[date].md.

    line 198
  7. Very specific:

    If verdict is PASS and the user agrees, ask: "May I update production/stage.txt to [to]?"

    line 215

“Common” is measured: kept by at least 6 in 10 of the 19 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

Gate Check

Phase advancement requires more than vibes. This skill collects every recent verdict artifact, cross-checks them for staleness and consistency, and emits a single PASS / CONCERNS / FAIL ruling. It is intentionally cross-system synthesis — therefore Opus tier.


Phase 1: Determine Source and Target Phase

Parse the argument:

  • Explicit form: prototype->preproduction, preproduction->production, production->polish, polish->release, release->live-ops.
  • --auto — read production/stage.txt and use the next standard transition.
  • No argument — ask the user.

Each gate has a different artifact bar (see Phase 3).


Phase 2: Inventory Recent Artifacts

Glob the latest of each artifact type:

  • production/qa/qa-signoff-*.md
  • production/qa/smoke-*.md
  • production/perf/perf-profile-*.md
  • production/security/security-audit-*.md
  • production/assets/asset-audit-*.md
  • production/balance/balance-check-*.md
  • production/localization/loc-report-*.md
  • docs/tech-debt.md
  • production/qa/soak/[date]/summary.md
  • production/releases/release-checklist-*.md
  • production/releases/launch-checklist-*.md
  • docs/architecture/ for ADR statuses
  • production/qa/bugs/ for open S1/S2 counts

For each, capture: path, last modified date, verdict.

Apply staleness rule: any artifact older than 14 days is suspect for a release-tier gate. Flag and surface.


Phase 3: Apply the Gate-Specific Bar

prototype -> preproduction

Required:

  • Architecture artifacts present (docs/architecture/master-architecture.md, ADRs, control manifest).
  • All MVP PRDs reviewed via /review-all-prds with PASS verdict.
  • No proposed ADRs blocking.

preproduction -> production

Required:

  • /test-setup complete and CI runs green.
  • Foundation epics done.
  • First sprint plan approved.

production -> polish

Required:

  • All Critical / Must-Have features implemented per PRD.
  • Latest QA sign-off APPROVED on the most recent sprint.
  • Open S1/S2 bugs == 0.
  • Tech debt register reviewed in last 30 days.

polish -> release

Required:

  • Latest perf profile HEALTHY.
  • Latest security audit RELEASE OK.
  • Soak test summary PASS.
  • Localization 100% on every target locale.
  • Asset audit CLEAN.
  • Balance check BALANCED.
  • Regression suite green.
  • Open S1/S2 == 0; open S3 reviewed.

release -> live-ops

Required:

  • /launch-checklist GO.
  • Day-one patch scoped or N/A.
  • Crash + analytics live and verified post-deploy.
  • On-call rotation defined.
  • Rollback plan documented.

Phase 4: Cross-Verdict Consistency

Look for contradictions between artifacts:

  • QA sign-off says "no S1/S2" but production/qa/bugs/ shows open S1 -> contradiction.
  • Perf profile HEALTHY but soak summary lists memory growth -> recheck.
  • Security RELEASE OK but tech debt register has a P0 deprecated-API with pending OS removal -> conflict.
  • Asset audit CLEAN but localization report has missing locales for store screenshots -> partial.

Surface every contradiction. None can be silently passed through.


Phase 5: Apply Severity Rules

Convert each artifact's status to a contribution:

  • Verdict PASS / GREEN / APPROVED -> +1.
  • Verdict WITH CONDITIONS / WARNINGS / NEEDS WORK -> 0 (note conditions).
  • Verdict FAIL / RED / NOT APPROVED / BLOCKING -> -1.
  • Stale (> 14 days) -> 0 with a note suggesting re-run.

Compute the gate result:

  • All required-for-this-gate artifacts at +1, no contradictions -> PASS.
  • Any -1 OR any required artifact missing OR any contradiction unresolved -> FAIL.
  • Otherwise (mix of +1 and 0 with no -1) -> CONCERNS.

Phase 6: Build Specific Blocker List

For FAIL verdicts, every -1 contributes a specific blocker line:

  • "QA sign-off NOT APPROVED — sprint-04 build failed manual QA on paywall flow. Re-run /team-qa after fix."
  • "Soak test FAIL — 12 MB/h leak after 4 hours. Open story to address before re-running soak."

For CONCERNS, every conditions/warnings entry contributes a recommended follow-up.


Phase 7: Render the Verdict

# Gate Check: [from] -> [to]
Date: [date]
Tier: [PASS / CONCERNS / FAIL]

## Artifact Inventory
| Artifact | Date | Verdict | Stale? | Contribution |
|----------|------|---------|--------|--------------|
| QA sign-off (sprint-04) | 2026-04-29 | APPROVED | no | +1 |
| Perf profile | 2026-05-01 | HEALTHY | no | +1 |
| Security audit | 2026-04-12 | RELEASE WITH PLAN | borderline | 0 |
| Soak summary | (none) | MISSING | — | -1 |

## Cross-Verdict Conflicts
- [list, or "None"]

## Blockers
- [each -1 contribution as a one-line action]

## Concerns
- [each 0 contribution as a recommended follow-up]

## Required Artifacts Missing
- [list]

## Verdict: [PASS / CONCERNS / FAIL]

Rationale: [synthesis paragraph]

## Recommended Path
- [if PASS] Update `production/stage.txt` to [to] and proceed.
- [if CONCERNS] Either accept conditions or address top 1-2 concerns.
- [if FAIL] Address listed blockers in priority order; re-run gate
  after each.

Ask before writing to production/gates/gate-[from-to]-[date].md.


Phase 8: Update Session State

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

## Gate Check — [date]
- Direction: [from] -> [to]
- Verdict: [PASS / CONCERNS / FAIL]
- Blockers: [count]
- Path: [path]
- Next: [advance | resolve and re-run]

If verdict is PASS and the user agrees, ask: "May I update production/stage.txt to [to]?"


Quality Gates / PASS-FAIL

This skill IS the gate. The verdict is binding for phase advancement.


Examples

Example 1 — polish -> release: QA APPROVED, perf HEALTHY, security RELEASE OK, soak PASS, localization 100%, asset CLEAN, balance BALANCED, regression GREEN. Verdict: PASS. Stage updated to release.

Example 2 — production -> polish: QA APPROVED but 3 S2 bugs open in production/qa/bugs/. Verdict: FAIL with one blocker: "Resolve open S2 bugs (BUG-104, BUG-109, BUG-117) or document explicit deferral."


Next Steps

  • PASS -> proceed with stage transition.
  • CONCERNS -> resolve top concerns, re-run if release-tier.
  • FAIL -> drive blockers to closure; re-run.

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.