agentleFS
ShiningPlagueShiningPlague/shiningplague-studio/skills/architecture-review/SKILL.md

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

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.
  • lilangMaxSkill

    soak-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

  • lilangMaxSkill

    team-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

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
  1. Very specific:

    Before reading any full document, use Grep to extract ## Summary sections:

    line 48
  2. Very specific:

    All ADRs at docs/adr/[0-9]*.md (docs/architecture/ holds the blueprint + manifest, never ADRs)

    line 74
  3. Very specific:

    All files in docs/engine-reference/<engine>/modules/

    line 82
  4. Very specific:

    Read docs/architecture/tr-registry.yaml if it exists.

    line 101
  5. Distinctive:

    Check if decision text implicitly covers the requirement

    line 130
  6. Distinctive:

    Do all ADRs that mention an engine version agree on the same version?

    line 197
  7. Distinctive:

    Verify against module reference docs

    line 202
  8. Distinctive:

    Check for contradictory assumptions

    line 203
  9. Distinctive:

    If none: "No GDD revision flags — all GDD assumptions consistent with verified engine behaviour."

    line 238
  10. Distinctive:

    Ask: "Should I flag these GDDs for revision in the systems index?" If yes, update Status to exact string "Needs Revision" (no …

    line 240
  11. Very specific:

    If docs/architecture/architecture.md exists, validate:

    line 246
  12. Very specific:

    Every system from systems-index.md appears in architecture layers

    line 247
  13. Distinctive:

    No orphaned architecture (systems in arch doc with no GDD)

    line 250
  14. Distinctive:

    Ask via AskUserQuestion: "Review complete.

    line 298
  15. Very specific:

    Ask: "May I update docs/architecture/tr-registry.yaml with new requirement IDs?"

    line 309
  16. Very specific:

    Update last_updated and version at top

    line 316
  17. Distinctive:

    Only append CONFLICT entries (not GAP).

    line 331
  18. Very specific:

    After writing approved files, silently append to production/session-state/active.md:

    line 335
  19. Very specific:

    Update registry at data/_schemas/system_registry.json:

    line 347
  20. Distinctive:

    If any spawned agent returns BLOCKED, errors, or fails:

    line 367
  21. Distinctive:

    Assess dependencies: If output required by later phase, don't proceed without user input

    line 369

“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.

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

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 ADR
  • consistency: Cross-ADR conflict detection only
  • engine: Engine compatibility audit only
  • single-gdd [path]: Review for one specific GDD
  • rtm: 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 3

Also 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.md
  • docs/engine-reference/<engine>/breaking-changes.md
  • docs/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 + dependencies
  • docs/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:

  1. Exact/near match → reuse existing TR-ID. Update requirement text only if wording changed; add revised: [date].
  2. No match → assign new ID: TR-[system]-NNN, sequence + 1.
  3. 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:

CategoryExample
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:

  1. Read ## GDD Requirements Addressed section
  2. Check if decision text implicitly covers the requirement
  3. Mark coverage status:
StatusMeaning
✅ CoveredAn ADR explicitly addresses this requirement
⚠️ PartialAn ADR partially covers this, or coverage is ambiguous
❌ GapNo 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

  1. Collect Depends On fields from every ADR's "ADR Dependencies" section
  2. Topological sort → implementation order
  3. Flag unresolved dependencies (depends on Proposed ADR)
  4. Detect cycles → 🔴 DEPENDENCY CYCLE
  5. 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:

  1. Confirm or challenge each audit finding — specialists know engine nuances not in reference docs
  2. Identify engine-specific anti-patterns the audit may have missed (wrong Godot node type, etc.)
  3. 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:

  1. Post-cutoff API behaviour differs from training-data assumptions → check GDDs referencing the system
  2. Known engine limitations in ADRs → check GDDs designing mechanics around affected features
  3. 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.md appears 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 requirement text + revised date for entries whose GDD wording changed (ID stays same)
  • Mark status: deprecated for entries whose GDD requirement no longer exists (confirm before)
  • Never renumber or delete
  • Update last_updated and version at 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].md

Update registry at data/_schemas/system_registry.json:

  • For each new ADR referenced, ensure documentation_stack.adr_index.adrs[] has the entry

Phase 9: Handoff

  1. Immediate actions: Top 3 ADRs to create (highest-impact gaps first, Foundation before Feature)
  2. Gate guidance: "Run /gate-check pre-production when blocking issues resolved"
  3. Rerun trigger: "Re-run /architecture-review after 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:

  1. Surface immediately: Report "[AgentName]: BLOCKED — [reason]"
  2. Assess dependencies: If output required by later phase, don't proceed without user input
  3. Offer options via AskUserQuestion: skip / retry narrower scope / stop
  4. Always produce partial report — output whatever completed

Collaborative Protocol

  1. Read silently — don't narrate every file read
  2. Show the matrix — present full traceability matrix before asking
  3. Don't guess — if requirement is ambiguous, ask
  4. Ask before writing — always confirm
  5. Non-blocking — verdict is advisory; user decides whether to continue despite CONCERNS or FAIL

Project Path Reference

WhatCanonical path
ADRsdocs/adr/[0-9]*.md
Master GDD{{GDD_PATH}} + per-system docs/gdd/<system>.md
Systems registrydata/_schemas/system_registry.json (source) + docs/gdd/systems-index.md (generated view)
Engine referencedocs/engine-reference/<engine>/VERSION.md

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.