Claude Code Game Studios is a setup that organizes Claude Code into a coordinated game-development team of specialized AI agents. It supports game projects across design, programming, art, audio, narrative, quality assurance, and production, with skills and workflows for coordinating that work.
Getting it into your agent
It runs from inside its repository, so the clone comes first — what it calls does not travel with the file alone.
git clone --depth 1 https://github.com/Donchitos/Claude-Code-Game-Studiosnpx agentmods add skills/donchitos/claude-code-game-studios/test-evidence-reviewWrote 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/skills/donchitos/claude-code-game-studios/test-evidence-review)<a href="https://agentmods.dev/skills/donchitos/claude-code-game-studios/test-evidence-review"><img src="https://agentmods.dev/badge/skills/donchitos/claude-code-game-studios/test-evidence-review/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/skills/donchitos/claude-code-game-studios/test-evidence-review"><img src="https://agentmods.dev/badge/skills/donchitos/claude-code-game-studios/test-evidence-review.svg" alt="Reviewed on agentmods" width="80" height="20"></a>- NVIDIA SkillSpector warn
SkillSpector: 1 finding, up to medium
These are SkillSpector’s own severities. On a checked sample its high-severity flags on skills were ~96% false positives — a documented command, a public API, a “never do X” rule — so we show them as a caution to read, not a verdict. Why →
- medium Excessive Agency · line 7 Skill selects an external model or provider that may use a different account or billing plan than the operator expects. Undisclosed model switches can cause unexpected cost or quota consumption.Fix: Remove the model/provider override or disclose it prominently and require explicit operator approval before invoking an external coding CLI or billed model.
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 | $0.00038 | $0.04130 |
| Opus 5.5 | $0.00015 | $0.01652 |
| Sonnet 5.5 | $0.00008 | $0.00826 |
| Haiku 4.5 | $0.00004 | $0.00413 |
Grade A, and why
test-evidence-review scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 9d ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
Copies of this mod
3 near-identical copies found in the catalogue:
- test-evidence-review — 100% identical, 188 lines differ
- test-evidence-review — 100% identical, 188 lines differ
- test-evidence-review — 94% identical, 211 lines differ
How it starts
The opening of the file, as written. The whole thing — 361 lines — stays where its author put it; the contents beside it link to each section on GitHub.
!bash "${CLAUDE_SKILL_DIR}/../../hooks/yaml-helper.sh" resolve_config --keys automation,workflow,qa.level,testing.strict
Automation mode: Resolve modes.automation (project.local.yaml →
project.yaml → default collaborative). Every AskUserQuestion call and
every file write follows .claude/docs/automation-modes.md
(collaborative asks always · guided major-only · autonomous logs and proceeds;
automation_always_ask categories always prompt).
qa.level and testing.strict (resolved above) decide what this review may
call a blocker — the same rules /story-done closes stories by, so a BLOCKING
item here is one that would stop /story-done:
qa.level: minimalwaives tests. A Logic, Integration or Config/Data story with no test evidence (for Config/Data, no smoke report) isWAIVED (qa.level: minimal)— not MISSING — and raises no BLOCKING item; a test that does exist is still reviewed, but its gaps are ADVISORY. Visual/Feel and UI evidence (the retained screenshots, and the Visual/Feel sign-off) is never waived.- Gate level per story type: take
testing.strict.<type>from the resolved block (Logic→logic, Integration→integration, Visual/Feel→visual, UI→ui, Config/Data→config):true→ BLOCKING,false→ ADVISORY,unset→ a plain-booleantesting.strict: true|false(legacy form) applies to every type; else the Default Gate Level in.claude/docs/coding-standards.md(BLOCKING for Logic, Integration, Visual/Feel and UI; ADVISORY for Config/Data). Every "BLOCKING" below means that story type's gate level.
Test Evidence Review
/smoke-check verifies that test files exist and pass. This skill
goes further — it reviews the quality of those tests and evidence documents.
A test file that exists and passes may still leave critical behaviour uncovered.
A manual evidence doc that exists may lack the sign-offs required for closure.
Output: Summary report (in conversation) + optional production/qa/evidence-review-[date].md
When to run:
- Before QA hand-off sign-off (
/team-qaPhase 5) - On any story where test quality is in question
- As part of milestone review for Logic and Integration story quality audit
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.
- 9d ago Changed · +4 lines a7ab49090d30
- 29d ago Changed · +39 lines 4fda2284471e
- 1mo ago Changed dc4776268847
- 1mo ago Changed · +66 lines · -21 tokens per session 3989b573b0ec
- 1mo ago First seen · 252 lines · 59 tokens per session scan A 04cdf0b7e94a
test-evidence-review is a skill published in the GitHub repository Donchitos/Claude-Code-Game-Studios (28,478 stars, last pushed 10d ago), licensed MIT. It adds 38 tokens to every session and 4,130 once invoked, about $0.0002 per session on Opus 5.5. 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 2026-09-03.
Other skills, from other repositories
dare-execute
Executa uma task específica com implementação de código e testes. Use quando o usuário aprovar TASKS.md e quiser executar uma task. Implementa o código, roda testes (Ralph Loop) e valida até passar.
dare-bench
Roda o harness determinístico de fixtures de verificação (Fix·Rate, solve-rate, baseline). Mapeia o CLI dare bench.
dare-guard
Run the DARE security guard gate on artifacts (dare guard).
test-driven-development
Testing discipline that adapts to user tier and task type. Rigorous RED-GREEN-REFACTOR for critical code. Invisible testing for Builder-tier users. Pragmatic skip for trivial changes. The goal is confidence that the code works, not ceremony.
confidence-based-strictness
Use when the complexity or unfamiliarity of a task should increase guardrail enforcement. Scales verification depth based on how confident the system is in the correctness of the generated code.
quality-gates
Graduated pre-commit quality checks that scale with project complexity and user tier. Level 0 for Builder-tier small projects, up to Level 4 for enterprise codebases. Thresholds configurable in architecture.md.