architecture-review
Skill0 starsChanged 4 months agoMITView on GitHub
Use when validating architecture completeness against GDDs, checking ADR coverage gaps, detecting cross-ADR conflicts, verifying engine compatibility, or when the designer says 'review architecture', 'check ADR coverage', or 'are our ADRs complete?' Produces PASS/CONCERNS/FAIL verdict. ShiningPlague-adopted: reads ADRs from docs/adr/, integrates with system_registry.json + tr-registry.yaml, adds engine-specialist consult.
78 other projects use a version of this, 8 of them edited.
Adopted by
Most recently taken up by
- kinhluan/skillstheir own version, first seen Nov 8, 2026
- Smaui94/ClaudeSkillCollectiontheir own version, first seen Nov 8, 2026
- AndersonAndrad/skillstheir own version, first seen Nov 8, 2026
- frnkgns/Overgeared-Skillstheir own version, first seen Nov 8, 2026
- striderZA/OpenCodeGameStudiosedited, first seen Nov 4, 2026
- nledford/engineering-review-boardtheir own version, first seen Nov 3, 2026
Used with
Projects with it also use soak-test, team-live-ops 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.
lilangMaxSkillsoak-test
Generate a soak test protocol for extended play sessions. Defines what to observe, measure, and log during long play sessions to surface slow leaks, fatigue effects, and edge cases that only appear after sustained play. Primarily used in Polish and Release phases.
Also in 9 projects that use this
lilangMaxSkillteam-live-ops
Orchestrate the live-ops team for post-launch content planning: coordinates live-ops-designer, economy-designer, analytics-engineer, community-manager, writer, and narrative-director to design and plan a season, event, or live content update.
Also in 9 projects that use this
- +2 more from lilangMax
Other versions78 projects use a version of this.
Word for word · 0
None found.
Edited · 8
- Donchitos/Claude-Code-Game-Studios★ 27k13% same
- striderZA/OpenCodeGameStudios★ 11821% same
- dawn840705/codex-code-studios★ 1721% same
- lilangMax/ClaudeCodeGameStudios★ 822% same
- hongyuanc/codex-game-studios★ 518% same
- dawn840705/claude-code-studios★ 122% same
- HunLuanZhiZhu/Game-Studios-Plugin★ 120% same
- venus2099/codex-game-studios★ 114% same
Wrote their own version · 70
- langwatch/langwatch★ 5ksame name
- ghuntley/loom★ 1.7ksame name
- microsoft/hve-core★ 1.6ksame name
- DataDog/dd-trace-js★ 840same name
- decebals/claude-code-java★ 763same name
- borgbase/vykar★ 722same name
- appdevforall/CodeOnTheGo★ 586same name
- getsentry/warden★ 423same name
- owainlewis/blueprint★ 423same name
- openbootdotdev/openboot★ 277same name
60 more not listed
How matches are found. Word for word: the exact same text. Edited: shares passages with this file (“99% same” means 99% of its five-word phrases also appear here; 10% or more counts). Their own version: similar meaning, or, weakest, the same name. We can’t tell which project had it first. Projects are counted among the 200 most-starred candidates, and mentions are links to the repository we found, up to 50 per site, so both counts are a floor.
In this repository68 other files it gives its agents.
- adoptSkill
- architecture-decisionSkill
- art-bibleSkill
- asset-specSkill
- brainstormingSkill
- bug-reportSkill
- bug-triageSkill
- code-reviewSkill
- consistency-checkSkill
- content-auditSkill
- create-architectureSkill
- create-control-manifestSkill
- create-epicsSkill
- create-storiesSkill
- day-one-patchSkill
- design-reviewSkill
- design-systemSkill
- dev-storySkill
- dispatching-parallel-agentsSkill
- estimateSkill
- executing-plansSkill
- finishing-a-development-branchSkill
- gate-checkSkill
- goal-add-minorSkill
- goal-checkSkill
- goal-setSkill
- godot-engineSkill
- helpSkill
- map-systemsSkill
- narrative-designerSkill
- project-stage-detectSkill
- propagate-design-changeSkill
- qa-planSkill
- quick-designSkill
- red-flag-scanSkill
- regression-suiteSkill
- retrospectiveSkill
- review-all-gddsSkill
- safety-pushSkill
- scope-checkSkill
- security-auditSkill
- session-closeSkill
- setup-engineSkill
- sprint-planSkill
- startSkill
- story-doneSkill
- story-readinessSkill
- subagent-driven-developmentSkill
- systematic-debuggingSkill
- team-artSkill
- team-audioSkill
- team-combatSkill
- team-economySkill
- team-infraSkill
- team-levelSkill
- team-marketingSkill
- team-narrativeSkill
- team-polishSkill
- team-releaseSkill
- team-securitySkill
- The first 60 shown.
The file
Show its rules and text
We split this skill into the rules it gives an agent
21 rules. 0 are common, 11 are distinctive, and 10 are very specific to this project.
0
Common
Most of the 70 edited versions we read keep these.
11
Distinctive
This author's own choices. Worth a look.
10
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Very specific: line 48
Before reading any full document, use Grep to extract ## Summary sections:
- Very specific: line 74
All ADRs at docs/adr/[0-9]*.md (docs/architecture/ holds the blueprint + manifest, never ADRs)
- Very specific: line 82
All files in docs/engine-reference/<engine>/modules/
- Very specific: line 101
Read docs/architecture/tr-registry.yaml if it exists.
- Distinctive: line 130
Check if decision text implicitly covers the requirement
- Distinctive: line 197
Do all ADRs that mention an engine version agree on the same version?
- Distinctive: line 202
Verify against module reference docs
- Distinctive: line 203
Check for contradictory assumptions
- Distinctive: line 238
If none: "No GDD revision flags — all GDD assumptions consistent with verified engine behaviour."
- Distinctive: line 240
Ask: "Should I flag these GDDs for revision in the systems index?" If yes, update Status to exact string "Needs Revision" (no …
- Very specific: line 246
If docs/architecture/architecture.md exists, validate:
- Very specific: line 247
Every system from systems-index.md appears in architecture layers
- Distinctive: line 250
No orphaned architecture (systems in arch doc with no GDD)
- Distinctive: line 298
Ask via AskUserQuestion: "Review complete.
- Very specific: line 309
Ask: "May I update docs/architecture/tr-registry.yaml with new requirement IDs?"
- Very specific: line 316
Update last_updated and version at top
- Distinctive: line 331
Only append CONFLICT entries (not GAP).
- Very specific: line 335
After writing approved files, silently append to production/session-state/active.md:
- Very specific: line 347
Update registry at data/_schemas/system_registry.json:
- Distinctive: line 367
If any spawned agent returns BLOCKED, errors, or fails:
- Distinctive: line 369
Assess dependencies: If output required by later phase, don't proceed without user input
“Common” is measured: kept by at least 6 in 10 of the 70 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.
Files
Architecture Review
🌱 ShiningPlague-adopted. Originally Donchitos; battle-tested on a shipped Godot project. Studio ADRs live at
docs/adr/[0-9]*.md. Path conventions corrected throughout + project sources added (data/_schemas/system_registry.json,tr-registry.yaml, engine-specialist Agent dispatch).
Validates that the complete body of architectural decisions covers all game design requirements, is internally consistent, and correctly targets the project's pinned engine version. Quality gate between Technical Setup and Pre-Production.
If an artifact named here is absent: say so plainly in one line, skip that step, and continue. Never invent the file to satisfy a checklist, and never fail a close because an optional artifact was never created.
Argument Modes
- No argument /
full: Full review — all phases coverage: Traceability only — which GDD requirements have no ADRconsistency: Cross-ADR conflict detection onlyengine: Engine compatibility audit onlysingle-gdd [path]: Review for one specific GDDrtm: Full Requirements Traceability Matrix (GDD → ADR → Story → Test)
Phase 1: Load Everything
Phase 1a — Summary Scan (fast, low tokens)
Before reading any full document, use Grep to extract ## Summary sections:
Grep pattern="## Summary" glob="docs/gdd/*.md" output_mode="content" -A 4
Grep pattern="## Summary" glob="docs/adr/[0-9]*.md" output_mode="content" -A 3Also scan master GDD for system sections:
Grep pattern="^## [0-9]" path="{{GDD_PATH}}" output_mode="content"For single-gdd [path] mode: use the target GDD's summary to identify which ADRs reference the same system, then full-read only those ADRs. Skip full-reading unrelated GDDs.
For engine mode: only full-read ADRs — GDDs not needed.
For coverage or full mode: proceed to full-read everything below.
Phase 1b — Full Document Load
Design Documents:
{{GDD_PATH}}— master GDD (all system designs until per-system split)- Any per-system GDDs at
docs/gdd/<system>.md docs/gdd/systems-index.md— auto-generated from registry (authoritative system list)
Architecture Documents:
- All ADRs at
docs/adr/[0-9]*.md(docs/architecture/holds the blueprint + manifest, never ADRs) docs/architecture/architecture.md(if exists)docs/architecture/control-manifest.md
Engine Reference (project-owned and optional -- /setup-engine step 2 creates it):
docs/engine-reference/<engine>/VERSION.mddocs/engine-reference/<engine>/breaking-changes.mddocs/engine-reference/<engine>/deprecated-apis.md- All files in
docs/engine-reference/<engine>/modules/
Project Standards:
.claude/docs/technical-preferences.md
Project Additional Sources:
data/_schemas/system_registry.json— authoritative system status + dependenciesdocs/architecture/tr-registry.yaml— existing requirement IDs (preserve across runs)
Report: "Loaded [N] GDDs, [M] ADRs, [P] systems in registry, engine: Godot 4.6.1."
Also read docs/consistency-failures.md if it exists; surface recurring patterns as a "Known conflict-prone areas" note at the top of Phase 4 output.
Phase 2: Extract Technical Requirements from Every GDD
Pre-load the TR Registry
Read docs/architecture/tr-registry.yaml if it exists. Index existing entries by id and normalised requirement text. This prevents ID renumbering across review runs.
For each requirement extracted:
- Exact/near match → reuse existing TR-ID. Update
requirementtext only if wording changed; addrevised: [date]. - No match → assign new ID:
TR-[system]-NNN, sequence + 1. - Ambiguous → ask user: "Same as TR-[system]-NNN, or new?"
Skip entries with status: deprecated.
For each GDD, extract technical requirements — things architecture must provide:
| Category | Example |
| Data structures | "Each entity has health, max health, status effects" |
| Performance constraints | "Collision detection must run at 60fps with 200 entities" |
| Engine capability | "Inverse kinematics for character animation" |
| Cross-system communication | "Damage system notifies UI and audio simultaneously" |
| State persistence | "Player progress persists between sessions" |
| Threading/timing | "AI decisions happen off the main thread" |
| Platform requirements | "Supports keyboard, gamepad, touch" |
Cross-reference system_registry.json system status (active/wip/partial/phasing_out) to mark requirement status.
Phase 3: Build the Traceability Matrix
For each TR, search all ADRs:
- Read
## GDD Requirements Addressedsection - Check if decision text implicitly covers the requirement
- Mark coverage status:
| Status | Meaning |
| ✅ Covered | An ADR explicitly addresses this requirement |
| ⚠️ Partial | An ADR partially covers this, or coverage is ambiguous |
| ❌ Gap | No ADR addresses this requirement |
Build the full matrix:
| TR-ID | GDD | System | Requirement | ADR Coverage | Status |
|-------|-----|--------|-------------|--------------|--------|
| TR-combat-001 | combat.md | Combat | Hitbox detection < 1 frame | ADR-0003 | ✅ |
| TR-combat-002 | combat.md | Combat | Combo window timing | — | ❌ GAP |Count totals: X covered, Y partial, Z gaps.
Phase 3b: Story + Test Linkage (rtm mode only)
Glob production/epics/**/*.md (excluding EPIC.md index). For each story: extract TR-ID, file path, title, Status, ## Test Evidence test file path.
Glob tests/**/* and tools/**_check.gd. Build index: system → [test file paths].
For each test path from stories, confirm via Glob whether file exists. Note MISSING if stated but not found.
Extended matrix adds Story + Test File + Test Status (COVERED / MISSING / NONE / NO STORY) columns.
Phase 4: Cross-ADR Conflict Detection
Compare every ADR against every other for:
- Data ownership conflict — two ADRs claim same data
- Integration contract conflict — incompatible interface assumptions
- Performance budget conflict — combined exceeds frame budget
- Dependency cycle — A requires B requires A
- Architecture pattern conflict — signals vs direct calls to same system
- State management conflict — dual authority over same game state
For each conflict, output:
## Conflict: [ADR-NNNN] vs [ADR-MMMM]
Type: [Data ownership / Integration / Performance / Dependency / Pattern / State]
ADR-NNNN claims: [...]
ADR-MMMM claims: [...]
Impact: [what breaks if both implemented as written]
Resolution options:
1. [Option A]
2. [Option B]ADR Dependency Ordering
- Collect
Depends Onfields from every ADR's "ADR Dependencies" section - Topological sort → implementation order
- Flag unresolved dependencies (depends on Proposed ADR)
- Detect cycles → 🔴 DEPENDENCY CYCLE
- Output recommended implementation order (Foundation → Core → Feature)
Phase 5: Engine Compatibility Cross-Check
Version Consistency
- Do all ADRs that mention an engine version agree on the same version?
- Flag any ADR written for an older engine version as potentially stale.
Post-Cutoff API Consistency
- Collect "Post-Cutoff APIs Used" fields from all ADRs
- Verify against module reference docs
- Check for contradictory assumptions
Deprecated API Check
- Grep all ADRs for API names listed in
deprecated-apis.md - Flag any ADR referencing a deprecated API
Missing Engine Compatibility Sections
- List ADRs missing the section entirely (blind spots)
Engine Specialist Consultation
Spawn godot-specialist (or primary specialist from .claude/docs/technical-preferences.md) via Agent for:
- Confirm or challenge each audit finding — specialists know engine nuances not in reference docs
- Identify engine-specific anti-patterns the audit may have missed (wrong Godot node type, etc.)
- Flag ADRs assuming engine behaviour different from pinned version
Incorporate findings under ### Engine Specialist Findings in Phase 5 output.
Phase 5b: Design Revision Flags (Architecture → GDD Feedback)
For each HIGH RISK engine finding from Phase 5, check whether any GDD makes an assumption that the verified engine reality contradicts.
Cases to check:
- Post-cutoff API behaviour differs from training-data assumptions → check GDDs referencing the system
- Known engine limitations in ADRs → check GDDs designing mechanics around affected features
- Deprecated API conflicts → check GDDs assuming deprecated API behaviour
### GDD Revision Flags (Architecture → Design Feedback)
| GDD | Assumption | Reality (from ADR/engine-reference) | Action |
|-----|-----------|--------------------------------------|--------|If none: "No GDD revision flags — all GDD assumptions consistent with verified engine behaviour."
Ask: "Should I flag these GDDs for revision in the systems index?" If yes, update Status to exact string "Needs Revision" (no parentheticals — other skills match the exact string).
Phase 6: Architecture Document Coverage
If docs/architecture/architecture.md exists, validate:
- Every system from
systems-index.mdappears in architecture layers - Data flow covers all cross-system communication from GDDs
- API boundaries support all integration requirements
- No orphaned architecture (systems in arch doc with no GDD)
Phase 7: Output the Review Report
## Architecture Review Report — {{PROJECT_NAME}}
Date: [date]
Engine: Godot 4.6.1
GDDs Reviewed: [N]
ADRs Reviewed: [M]
### Traceability Summary
Total requirements: [N] — ✅ Covered: [X] — ⚠️ Partial: [Y] — ❌ Gaps: [Z]
### Coverage Gaps (no ADR exists)
[Per gap: TR-ID, GDD, system, requirement, suggested ADR title, engine risk]
### Cross-ADR Conflicts
[All conflicts from Phase 4]
### ADR Dependency Order
[Topologically sorted, with unresolved deps and cycles]
### GDD Revision Flags
[From Phase 5b, or "None"]
### Engine Compatibility Issues
[From Phase 5]
### Architecture Document Coverage
[From Phase 6]
### Verdict: [PASS / CONCERNS / FAIL]
PASS: All requirements covered, no conflicts, engine consistent
CONCERNS: Some gaps/partial, no blocking conflicts
FAIL: Foundation/Core gaps, or blocking conflicts
### Blocking Issues (FAIL only)
### Required ADRs (prioritised, Foundation first)Phase 8: Write and Update
Ask via AskUserQuestion: "Review complete. What would you like to write?"
- [A] Write all three files (review report + traceability index + TR registry)
- [B] Write review report only at
docs/architecture/architecture-review-[date].md - [C] Don't write — review findings first
RTM Output (rtm mode only)
Additionally ask: "May I write full RTM to docs/architecture/requirements-traceability.md?"
TR Registry Update
Ask: "May I update docs/architecture/tr-registry.yaml with new requirement IDs?"
If yes:
- Append new TR-IDs not in registry
- Update
requirementtext +reviseddate for entries whose GDD wording changed (ID stays same) - Mark
status: deprecatedfor entries whose GDD requirement no longer exists (confirm before) - Never renumber or delete
- Update
last_updatedandversionat top
Reflexion Log Update
Append any 🔴 CONFLICT entries from Phase 4 to docs/consistency-failures.md (if file exists, append only):
### [YYYY-MM-DD] — /architecture-review — 🔴 CONFLICT
**Domain**: Architecture / [domain]
**Documents involved**: [ADR-NNNN] vs [ADR-MMMM]
**What happened**: [conflict]
**Resolution**: [how resolved]
**Pattern**: [generalised lesson]Only append CONFLICT entries (not GAP). Don't create file if missing.
Session State Update
After writing approved files, silently append to production/session-state/active.md:
## Session Extract — /architecture-review [date]
- Verdict: [PASS / CONCERNS / FAIL]
- Requirements: [N] total — [X] covered, [Y] partial, [Z] gaps
- New TR-IDs registered: [N, or "None"]
- GDD revision flags: [comma-separated GDD names, or "None"]
- Top ADR gaps: [top 3 gap titles, or "None"]
- Report: docs/architecture/architecture-review-[date].mdUpdate registry at data/_schemas/system_registry.json:
- For each new ADR referenced, ensure
documentation_stack.adr_index.adrs[]has the entry
Phase 9: Handoff
- Immediate actions: Top 3 ADRs to create (highest-impact gaps first, Foundation before Feature)
- Gate guidance: "Run
/gate-check pre-productionwhen blocking issues resolved" - Rerun trigger: "Re-run
/architecture-reviewafter each new ADR"
Close with AskUserQuestion: "Architecture review complete. What would you like to do next?"
- [A] Write a missing ADR — open fresh session, run
/architecture-decision [system] - [B] Run
/gate-check pre-production— if all blocking gaps resolved - [C] Stop here for this session
Error Recovery Protocol
If any spawned agent returns BLOCKED, errors, or fails:
- Surface immediately: Report "[AgentName]: BLOCKED — [reason]"
- Assess dependencies: If output required by later phase, don't proceed without user input
- Offer options via AskUserQuestion: skip / retry narrower scope / stop
- Always produce partial report — output whatever completed
Collaborative Protocol
- Read silently — don't narrate every file read
- Show the matrix — present full traceability matrix before asking
- Don't guess — if requirement is ambiguous, ask
- Ask before writing — always confirm
- Non-blocking — verdict is advisory; user decides whether to continue despite CONCERNS or FAIL
Project Path Reference
| What | Canonical path |
| ADRs | docs/adr/[0-9]*.md |
| Master GDD | {{GDD_PATH}} + per-system docs/gdd/<system>.md |
| Systems registry | data/_schemas/system_registry.json (source) + docs/gdd/systems-index.md (generated view) |
| Engine reference | docs/engine-reference/<engine>/VERSION.md |
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.

