Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
npx skills add DY-2026/GameDesignOS --skill game-experience-density-optimizergit clone --depth 1 https://github.com/DY-2026/GameDesignOSWrote 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/dy-2026/gamedesignos/game-experience-density-optimizer)<a href="https://agentmods.dev/skills/dy-2026/gamedesignos/game-experience-density-optimizer"><img src="https://agentmods.dev/badge/skills/dy-2026/gamedesignos/game-experience-density-optimizer/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/dy-2026/gamedesignos/game-experience-density-optimizer"><img src="https://agentmods.dev/badge/skills/dy-2026/gamedesignos/game-experience-density-optimizer.svg" alt="Reviewed on agentmods" width="80" height="20"></a>- NVIDIA SkillSpector pass
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 | $0.00124 | $0.03852 |
| Opus 5 | $0.00062 | $0.01926 |
| Sonnet 5 | $0.00025 | $0.00770 |
| Haiku 4.5 | $0.00012 | $0.00385 |
Grade A, and why
game-experience-density-optimizer 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 10d 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 — 230 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Game Experience Density Optimizer
Copyright (c) 2026 Paranoia. Licensed under the MIT License.
Mission
把模糊的游戏体验问题编译成可上线、可埋点、可复盘、可回滚的 ED 实验包。
这里的 ED 是 Experience Density / 体验浓度。中文统一叫“体验浓度”,不要另造概念名。它不是科学量表,也不是留存玄学;输出必须默认标注 theory_status: design_hypothesis,并把结论绑定到证据等级、游戏形态、主旋钮、指标周期和回滚条件。
默认内部管线:
输入材料 -> 输出模式路由 -> Evidence Gate -> 游戏形态分流 -> 最佳刺激窗口 ->
ED 公式项定位 -> 主旋钮选择 -> 实验变体编译 -> 埋点/看板编译 ->
预注册决策门 -> 输出门检查
When To Use
用户讨论以下问题时触发本 skill:
- “体验浓度”、
ED / Experience Density、每分钟有多少有意义选择、首局太空、首个爆点太晚。 - 留存实验、D1/D3/D7、每日会话、回流 rehook、活动留存、老玩家钝化、中段疲劳。
- 单机总游戏时长、买断制完成率、Steam Demo 完成率、章节推进、核心循环到达率、重玩意愿。
- 反馈不爽、不清楚、不跟手、打击软、操控延迟、镜头/触觉/动作节拍问题。
- 氛围空、留白无质感、叙事停顿、信息太吵、认知负荷高。
- 最佳刺激、低刺激无聊、过载无聊、习惯化、半熟半新、可控惊讶。
- FEP/free-energy、预测误差、Markov blanket、玩家和游戏的输入输出边界。
- 一周 A/B 测试、埋点字典、看板字段、预注册规则、回滚/Kill 条件。
不要用于只有一句创意、还没有核心循环的任务;先用 game-concept-architect。不要把截图、PV 或商店页直接当真实节奏证据;先用 game-experience-analyzer 建证据层。不要设计暗黑模式、误导奖励、焦虑红点、虚假倒计时、付费压力或不可逆损失伪装。
Mode Router
先判断输出模式,再决定交付深度。强 skill 的默认不是写大报告,而是给当前场景刚好够用的结果。
| mode | 触发 | 输出密度 |
|---|---|---|
quick_ed_triage |
用户只给一句体验问题,或明确要快速判断 | 1 个边界判断、1 个刺激窗口、1 个主旋钮、2 个最小改动、3 个验证指标、1 个回滚条件 |
weekly_ab_plan |
用户问怎么改、怎么测、本周怎么做、A/B 测试、留存实验、实验方案 | A/B 或 A/B/C/D 变体、埋点、看板、决策门、owner、回滚 |
instrumentation_plan |
用户重点问埋点、看板、指标口径、数据接线 | 事件字典、字段、触发时机、过滤器、数据质量门、隐私边界 |
review_and_decide |
用户提供实验结果、指标变化、复盘材料 | 先查负向门和数据质量,再决定 amplify / iterate / observe / rollback / kill |
full_client_delivery |
用户要求客户交付、团队方案、完整文档、正式报告 | 展开完整 19 模块,附 handoff checklist、QA、风险门 |
schema_json |
用户要求 agent 消费、自动化验证、结构化输出 | 输出符合 templates/experiment-plan.schema.json 的 JSON,保留证据和 unknown 字段 |
如果用户没有说明模式:一句话问题默认 quick_ed_triage;出现“本周、实验、A/B、怎么测、留存方案”默认 weekly_ab_plan;出现“完整、交付、客户、团队评审”默认 full_client_delivery。
Hard Gates
所有输出必须经过这些门:
evidence_gate:先声明evidence_level、evidence_status、允许结论、禁止结论、置信度、缺失证据和混淆风险。读取references/evidence-gate.zh-CN.md。metric_horizon_gate:先判断game_metric_model:premium_single_player、mobile_liveops、hybrid或unknown。单机/买断制默认总旅程指标;手游/liveops 才默认 D1/D7。stimulation_window_gate:先判断最佳刺激窗口和无聊类型。无聊不自动等于刺激不足。density_formula_gate:把问题落到CLP、SF、EB、AR、MD/min,并说明为什么。one_primary_lever_gate:每个变体只能有一个主旋钮,最多一个不影响归因的辅助动作。instrumentation_gate:没有埋点/看板/复盘口径的方案不能说已可验证。decision_rule_gate:成功、观察、回滚、Kill 条件必须在实验前写死。ethics_gate:不得用暗黑模式或纯数值膨胀伪装体验优化。output_density_gate:不要在quick_ed_triage里输出完整 19 模块;不要在full_client_delivery里省略关键风险门。
What ships with it
33 files beside SKILL.md in the same directory: the scripts, references and assets a skill reads on demand. Not counted in the per-session cost; read them before you install if any of them is executable.
- agents/openai.yaml 1.1 KB
- evals/behavior_evals.json 410 B
- evals/evals.json 6.7 KB
- evals/negative_cases.md 5.1 KB
- evals/rubric.yaml 3.9 KB
- evals/synthetic_outputs.json 3.2 KB
- examples/synthetic-hybrid-conflict-review.md 2.7 KB
- examples/synthetic-premium-demo-completion-ed-plan.md 3.8 KB
- examples/synthetic-survivors-first-session-ed-plan.md 3.6 KB
- README.md 11 KB
- references/density-diagnosis-workflow.zh-CN.md 8.0 KB
- references/density-formula.zh-CN.md 9.2 KB
- references/ed-framework.zh-CN.md 7.2 KB
- references/ed-handoff-contract.md 386 B
- references/evidence-gate.zh-CN.md 4.8 KB
- references/flow-sdt-experience-gates.zh-CN.md 4.5 KB
- references/free-energy-markov-blanket-lens.zh-CN.md 9.6 KB
- references/interaction-prediction-lens.zh-CN.md 4.5 KB
- references/lever-playbook.zh-CN.md 8.7 KB
- references/metric-horizon-by-game-model.zh-CN.md 4.9 KB
- references/optimal-stimulation-window.zh-CN.md 9.2 KB
- references/retention-risk-gates.zh-CN.md 5.1 KB
- references/telemetry-metric-dictionary.zh-CN.md 10 KB
- references/theory-source-map.zh-CN.md 4.1 KB
- references/weekly-experiment-sop.zh-CN.md 4.1 KB
- templates/dashboard-spec.md 3.3 KB
- templates/ed-scorecard.md 1.6 KB
- templates/experiment-intake.md 4.5 KB
- templates/experiment-plan.schema.json 13 KB
- templates/instrumentation-dictionary.md 4.4 KB
- templates/variant-matrix.md 2.6 KB
- templates/weekly-ed-experiment-plan.md 9.3 KB
- templates/weekly-review.md 1.7 KB
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.
- 10d ago First seen · 230 lines · 124 tokens per session scan A dc672f33e479
game-experience-density-optimizer is a skill published in the GitHub repository DY-2026/GameDesignOS (383 stars, last pushed 23d ago), licensed MIT. It adds 124 tokens to every session and 3,852 once invoked, about $0.0006 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 skills, from other repositories
aigd-ui-capture
AIGD · UI screenshot → UI DSL (knowledge-base capture / tool 1). Use this when you want to turn a game UI screenshot (a competitor's or your own) into structured md (type/hierarchy/geometry/interaction/state), so uirender.py can restore it to a wireframe and it can be accumulated into the patterns/UI-paradigm…
aigd
AIGD (AI Game Design) main entry / orchestrator. Use this when the user wants to start a game project from scratch, is unsure which phase to use, wants to survey or advance the entire development pipeline, or needs to maintain the project spine (project charter / manifest). This skill only does "read the spine → judge…
aigd-handoff
AIGD Phase 4 · finalize and generate the handoff package. Use this when, after a system is finalized in playtesting, you want to generate the interface contract, acceptance cases, (as needed) art/client/server docs, and the designer-facing acceptance checklist. Produces platform-agnostic handoff artifacts that…
aigd-sync
AIGD cross-cutting action · sync-back integration (a non-linear "last step", triggered after every system's handoff). Use this when, once a system is settled, you need to update the global spec, the overall html prototype (integration demo), the implementation master guide (handoff-package entry point), or mark…
aigd-concept
AIGD Phase 1 · brainstorming & top-level design. Use this when you want to set the concept for a new game, fix the core loop / platform / target users, or split the game into a system list + dependency graph and start the project spine. Produces / revises the project charter + manifest (system list). Part of the aigd…
aigd-iterate
AIGD Phase 3 · playtest iteration. Use this when, after the user playtests the prototype, they want to optimize the rules or change the config table (can be run repeatedly). Only touches the "cheap" artifacts — rules / config / prototype — not the contract / acceptance (those are produced after finalization). Part of…