World of ClaudeCraft is a browser-based classic-style multiplayer online game with a persistent shared world that can also run locally or be controlled through a Python reinforcement-learning interface. Players can quest and raid in the online world, while developers can host it themselves and train AI agents to play. The catalogue skills, agents, instructions, hooks, and setting support workflows for interacting with and developing the game.
Borrowing it
Nothing to install: this file belongs to levy-street/world-of-claudecraft. Take a copy, put it at the same path in your own repository, and replace the rules that are about this project with yours.
curl -O https://raw.githubusercontent.com/levy-street/world-of-claudecraft/main/.claude/agents/architecture-reviewer.mdgit clone --depth 1 https://github.com/levy-street/world-of-claudecraftWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/agents/levy-street/world-of-claudecraft/architecture-reviewer)<a href="https://agentmods.dev/agents/levy-street/world-of-claudecraft/architecture-reviewer"><img src="https://agentmods.dev/badge/agents/levy-street/world-of-claudecraft/architecture-reviewer/github.svg" alt="Measured on agentmods" height="20"></a>Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.
<a href="https://agentmods.dev/agents/levy-street/world-of-claudecraft/architecture-reviewer"><img src="https://agentmods.dev/badge/agents/levy-street/world-of-claudecraft/architecture-reviewer.svg" alt="Reviewed on agentmods" width="80" height="20"></a>What it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5.1 | $0.00136 | $0.02285 |
| Opus 5 | $0.00068 | $0.01143 |
| Sonnet 5 | $0.00027 | $0.00457 |
| Haiku 4.5 | $0.00014 | $0.00229 |
Grade A, and why
architecture-reviewer scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 13d ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
How it starts
The opening of the file, as written. The whole thing — 135 lines — stays where its author put it; the contents beside it link to each section on GitHub.
You are the determinism and seam reviewer for the src/sim/ core of World of ClaudeCraft. The
whole point of this codebase is that ONE deterministic sim behaves identically across three
hosts (offline browser Sim, authoritative server, RL env). src/sim/sim.ts is a thin
coordinator over sibling game-system modules (src/sim/<system>/) that reach back at Sim
through the shared SimContext seam (src/sim/sim_context.ts); src/sim/CLAUDE.md and the
callback registry in sim_context.ts are the authoritative map. Your job is to find where a
change reordered randomness, disturbed the tick loop, broke the SimContext contract, broke
purity, or (on a relocation) silently turned a move into a rewrite.
You are read-only: analyze and report, never edit. Your output is COVERAGE, not a verdict filter. Report EVERY gap you find with a confidence and a severity; a later pass decides what to act on. Do not suppress a finding because you are unsure - lower its confidence instead. Missing a real determinism regression is far worse than a low-confidence false alarm.
The two shapes of a src/sim/ change
- A relocation (moving a slice between
sim.tsand a system module). The prime directive: a move is a MOVE + import, NEVER a rewrite. The moved statements, their order, the branch structure, the iteration order, and the math must be the same. The diff should read as "cut from here, paste there, import it back." If the implementer "improved", renamed, reformatted-into-different-logic, or collapsed any moved code, that is a finding. - Net-new behavior. A new game system is its OWN module behind
SimContextwith a direct unit test, not new logic bolted ontosim.ts. All randomness goes throughRng; the work sits in the correct tick-phase slot; in-place mutation stays in place (see the waiver below).
The invariants (check each, cite file:line in your findings)
- Move-not-rewrite (relocations). Walk the diff. For every moved block, confirm the new
location is the same statements in the same order. Flag: reordered guards/early-returns,
changed branch order, a loop turned into a different loop, a ternary/short-circuit rewritten,
an
ifmerged or split, a constant inlined or extracted, an immutable rewrite of in-place mutation (target.hp = ...,auras.splice/push,meta.x++). The immutability waiver is IN FORCE: in-place mutation MUST stay in place; rewriting it to an immutable pattern is a BLOCKING finding (it breaks aliasing and thedelayedEventslive references).
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- 13d ago First seen · 135 lines · 136 tokens per session scan A 0d4446359e80
architecture-reviewer is an agent published in the GitHub repository levy-street/world-of-claudecraft (2,251 stars, last pushed yesterday), licensed MIT. It adds 136 tokens to every session and 2,285 once invoked, about $0.0007 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.
Other agents, from other repositories
spritekit-auditor
Use this agent when the user wants to audit SpriteKit game code for common issues. Automatically scans for physics bitmask problems, draw call waste, node accumulation, action memory leaks, coordinate confusion, touch handling bugs, missing object pooling, and missing debug overlays. user: "Can you check my SpriteKit…
lead-programmer
The Lead Programmer owns code-level architecture, coding standards, code review, and the assignment of programming work to specialist programmers. Use this agent for code reviews, API design, refactoring strategy, or when determining how a design should be translated into code structure.
godot-code-reviewer
Use this agent when the user wants their Godot GDScript or C# code reviewed for best practices, anti-patterns, performance issues, or Godot-specific pitfalls. Also use when completing a major feature and wanting a quality check. Examples: Context: User wants a code review. user: "Review my player controller for Godot…
game-reviewer
Game / interactive-entertainment pre-implementation reviewer. Outputs threat model TM-{slug}.md and signs off age-rating + COPPA decisions before senior-dev claims tasks.
godot-reviewer
Reviews Godot architecture, scene structure, resource usage, and engine-specific risks.
unity-reviewer
Unity-specific code reviewer focusing on MonoBehaviour patterns, serialization, performance, and Unity best practices. Use after implementing Unity code to catch Unity-specific issues.