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.
git clone --depth 1 https://github.com/movebrickschi/harness-engineering-mcpWrote 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/rules/movebrickschi/harness-engineering-mcp/16-ai-efficiency)<a href="https://agentmods.dev/rules/movebrickschi/harness-engineering-mcp/16-ai-efficiency"><img src="https://agentmods.dev/badge/rules/movebrickschi/harness-engineering-mcp/16-ai-efficiency/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/rules/movebrickschi/harness-engineering-mcp/16-ai-efficiency"><img src="https://agentmods.dev/badge/rules/movebrickschi/harness-engineering-mcp/16-ai-efficiency.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 | $0.00693 | $0.00693 |
| Opus 5.5 | $0.00277 | $0.00277 |
| Sonnet 5.5 | $0.00139 | $0.00139 |
| Haiku 4.5 | $0.00069 | $0.00069 |
Grade A, and why
16-ai-efficiency 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 1mo 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.
What it actually says
AI Efficiency · 高效执行 & 省 Token 硬规则
与
assets/spec/AI_EFFICIENCY.md/assets/skills/ai-efficiency/SKILL.md配套。本文件是 Cursor / Claude Code 直接消费的硬规则版本——只列不可违反的最小集。
R1. 检索路径
- ✅ 知道符号名 → 先
Grep，禁止直接 Read 大文件找符号 - ✅ 知道文件名规律 → 先
Glob，再 Read - ❌ 一次 SemanticSearch 命中 ≥ 10 个文件 → 必须收紧查询，不要全读
R2. 读文件切片
- 单次 Read 必须带 offset/limit 当文件 > 500 行
- 永远不要 Read 整个
node_modules//dist//vendor/
R3. 并行调用
- 同一轮里独立的 ≥ 2 个工具调用必须一次性批发
- 顺序串行调用超过 4 次 = 违规
R4. Subagent 边界
- 检索范围 ≥ 3 文件 → 委派 subagent
- 主会话不读完整 fixture / 大日志 / 大 JSON dump，由 subagent 摘要后回报
R5. 修改原子化
- 一次
StrReplace改动 ≤ 100 行 diff - 跨文件大改 → 拆 ≥ N 次小 commit，每次均可独立通过
npm test --runRelatedTests
R6. 验证最小化
- 改完只跑被影响的测试子集（
--runRelatedTests/-pl module/ 单文件 pytest） - 全局
harness check --strict --run-tests仅在收尾跑一次
R7. 输出
- 每段
reply_message/ 主对话回复 < 2000 字符；超过即拆段或表格化 - 禁止重复粘贴 spec / rule 段落；用 URI 引用
R8. 失败处理
- 同一 bug ≥ 3 次尝试未修 → 切
systematic-debuggingskill，禁止「再试一下」式补丁 - 测试红了 → 优先根因，禁止 mock 掉测试断言来过
R9. 计划与执行
-
3 步任务 → 必须先出
writing-plansskill 的明文计划 - 用户拍板后才进入
executing-plans推进
R10. 缓存
- 不在每轮 prompt 里手抄 spec / rule / 项目 README
- 通过 MCP 资源 URI 引用，保持 prompt 前缀稳定以命中 prompt cache
任何项目可以在
.harness/config.json的modules.knowledge.ai_efficiency_strict中关闭部分硬规则，但默认全开。
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.
- 1mo ago First seen · 56 lines · 693 tokens per session scan A fcf6b8158e42
16-ai-efficiency is a cursor rule published in the GitHub repository movebrickschi/harness-engineering-mcp (2 stars, last pushed 4mo ago), licensed MIT. It adds 693 tokens to every session, about $0.0028 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 2026-08-31.
Other cursor rules, from other repositories
explorer-interface-brief
Use when asked to design the interface for an interactive 3D explorer, plan a UI for exploring a 3D model of a topic, or write an image-model prompt for a 3D learning app mock-up. Produces, from a topic, the UI design brief (audience, layout, panels, interactions, label style, colour palette, accessibility) and an…
reflection-basic
Basic reflection format for Level 2 Simple Enhancement tasks.
simplicity
Simplicity - smallest correct diff; reuse, platform, installed deps before new code; no speculative abstraction.
git-workflow
Rules for git operations, commit strategy, and GitHub CLI usage.
feedback-enhanced
Rules for an interactive feedback system that lets users and an AI agent discuss complex development work in real time. It includes guidance for recording decisions, tracking progress and handling detailed inputs such as code and diagrams.
kiss-ponytail
Ponytail, lazy senior dev mode. Always pick the simplest solution that works.