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 BottleYo/model-economy --skill cost-aware-developmentgit clone --depth 1 https://github.com/BottleYo/model-economyWrote 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/bottleyo/model-economy/cost-aware-development)<a href="https://agentmods.dev/skills/bottleyo/model-economy/cost-aware-development"><img src="https://agentmods.dev/badge/skills/bottleyo/model-economy/cost-aware-development/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/bottleyo/model-economy/cost-aware-development"><img src="https://agentmods.dev/badge/skills/bottleyo/model-economy/cost-aware-development.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 | $0.00040 | $0.01699 |
| Opus 5 | $0.00020 | $0.00849 |
| Sonnet 5 | $0.00008 | $0.00340 |
| Haiku 4.5 | $0.00004 | $0.00170 |
Grade A, and why
cost-aware-development 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.
How it starts
The opening of the file, as written. The whole thing — 66 lines — stays where its author put it; the contents beside it link to each section on GitHub.
成本感知开发
先按风险选择最低可胜任能力,再决定需要多少流程和角色。简单任务保持简单;成本优化不得绕过真实风险、安全边界或完成证据。
先识别运行模式
启动任务时先判断可选的六角色增强是否健康。若本地 CLI 可用,优先读取 model_economy.py status 的只读结果。CLI 不可用且当前环境没有任何 model-economy-* 自定义角色时,按核心模式运行;只要出现部分或全部同名角色但无法验证完整增强健康状态,就保守视为降级状态。不得因为文档中出现角色名称就假定角色、模型映射或身份验证已经启用。
core(核心模式):目录安装后四个 Skill 可直接运行,但没有本地六角色配置。简单、机械和标准任务均由主 agent 执行,继续应用原生质量门,不尝试调用不存在的自定义角色。enhanced(增强模式):六个角色文件、受管理哈希、模型映射和模板版本全部健康时,才按下方六角色规则编排。模型与角色身份仍然是未验证状态。degraded(降级状态):只要发现部分角色、损坏状态、哈希或映射冲突、过期模板,就失败关闭并明确要求修复或移除增强配置,不得静默当作核心模式。
核心模式遇到大型/高风险任务时,必须说明独立架构师和终审员的角色隔离条件未满足,并让用户选择安装增强模式,或明确批准“降低保障的单 agent 流程”。即使用户批准后者,也不得声称已经完成完整的 Model Economy 高风险流程。降级状态下同样不得作出这种完成声明。
唯一分类表
路由策略是任务分类、角色集合、条件门、完成门和 strong 上限的唯一机器事实源。按 classification_order 和 first_match 选择唯一类别;最后的 standard 是默认 fallback,不因文件数、请求篇幅或省 token 要求而降级风险。
跨模块不自动等于大型。只有新架构、广泛爆炸半径,或安全、隐私等高风险边界才进入大型/高风险。机械任务必须满足全部五项 economy 条件;任一条件不满足就使用 balanced。
原生质量内核
Model Economy 自带意图门、计划门、测试门和证据门,不依赖其他开发方法插件。开始任务时只判断各门是否命中;命中时才读取并执行质量门,不得为了形式启动未命中的流程。
目标存在实质歧义、高风险决策或未批准架构边界时,实施前必须取得可追溯审批。目标、约束和验收已经明确时,不重复要求设计审批。简单任务不写持久化计划;标准任务只使用与复杂度匹配的短计划;大型/高风险任务才要求持久化计划。测试强度匹配行为风险。声明完成前始终核对新鲜验证证据。
轻量工程能力
需要专项方法时,可以按需组合 domain-context、module-design 和 disposable-prototype。它们分别用于提炼已有领域约束、检查模块边界,以及用隔离的一次性实验回答具体未知;没有匹配触发条件时不要读取或执行。
这些叶子 Skill 不改变任务分类、模型选择、角色拓扑或 subagent 预算,不得接管计划、审查、提交或完成门。它们自身不启动 subagent;涉及实际产品代码的后续工作仍由本 Skill 分类和编排。详细触发条件以路由策略中的 lightweight_practices 为准。
Superpowers 兼容
安装、启用或发现 Superpowers 不构成启用完整 Superpowers 工作流的授权。默认由 Model Economy 独占任务分类、代理编排和质量门,不自动调用其完整 brainstorming、planning、subagent-driven-development、executing-plans 或 finishing 流程。
只有用户在当前任务中明确要求“完整 Superpowers”、“full Superpowers”或“Superpowers strict mode”时,才把本次任务的流程编排权交给 Superpowers。交权后 Model Economy 只提供模型与成本建议,不得启动 Model Economy 角色,也不得追加自己的计划、审查或验证流程。交权不跨任务保存;如果运行环境没有 Superpowers,应明确报告 strict mode 不可用。
用户只要求测试先行、系统化调试或完成前验证等单项方法,不等于授权完整 Superpowers,也不触发交权。
What ships with it
5 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.
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 First seen · 66 lines · 40 tokens per session scan A ee3bb1da31bd
cost-aware-development is a skill published in the GitHub repository BottleYo/model-economy (5 stars, last pushed 1mo ago), licensed MIT. It adds 40 tokens to every session and 1,699 once invoked, about $0.0002 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 2026-08-31.
Other skills, from other repositories
stss
Reduce defensive disclaimers, stacked hedging, and self-protective narration in proposals and decision-facing writing. Use when the user asks to rewrite or audit a proposal, plan, research contribution, executive summary, or similar text for directness. Do not use for ordinary code work or unrelated prose.
superloopy-loop
Use Superloopy's lightweight strict-evidence loop for Codex tasks that need durable progress, criteria, and artifact-backed completion.
inspect-codex-trajectory
Inspect or show a local Codex task trajectory, including turns, approximate model steps, assistant messages, reasoning summaries, tool calls, failures, compaction, token usage, and timing. Use when the user asks for a trajectory, execution trace, task timeline, slow-tool analysis, visual event ledger, or live…
marketplace-publish-validation
Corezoid marketplace pre-publication validator. Standalone skill — no external skill dependencies required. Use this skill whenever the user wants to publish, release, or submit a project or folder to the Corezoid marketplace, or asks to check if a project is ready to publish. Activate on phrases: "publish to…
release
A release workflow for the corezoid-ai-plugin project that prepares a new tagged software version through six defined phases.
corezoid-init
Corezoid environment setup specialist. Use when the user wants to connect to Corezoid, set up credentials, authenticate, pull a project, configure the environment, or start working with a Corezoid project for the first time. Activate when the user says "init", "setup", "connect to corezoid", "login", "pull workspace"…