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/TashanGKD/tashan-cursor-skillsWrote 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/tashangkd/tashan-cursor-skills/auto-experience-hook)<a href="https://agentmods.dev/rules/tashangkd/tashan-cursor-skills/auto-experience-hook"><img src="https://agentmods.dev/badge/rules/tashangkd/tashan-cursor-skills/auto-experience-hook.svg" alt="Measured on agentmods" 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 | $0.00121 | $0.07472 |
| Opus 5 | $0.00060 | $0.03736 |
| Sonnet 5 | $0.00024 | $0.01494 |
| Haiku 4.5 | $0.00012 | $0.00747 |
Grade A, and why
auto-experience-hook 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 8d 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 — 421 lines — stays where its author put it; the contents beside it link to each section on GitHub.
⚠️ [2026-03-23] 本规则已被 session-bootstrap.mdc 的 B1/B2/B4 步骤实质性覆盖
状态:功能性冗余(Deprecated)。本规则与 session-bootstrap.mdc 序列B 功能完全重叠:
- 本规则的 Step 1 = session-bootstrap B1(经验感知 → PENDING-EXPERIENCES)
- 本规则的 Step 1.5 = session-bootstrap B2(积压感知)
- 本规则的 Step 2/3 = session-bootstrap B3/B4(告知用户 + 写任务日志)
根因:session-bootstrap.mdc(alwaysApply:true,步骤序列形式,执行可靠性高)在 2026-03-21 创建后实质替代了本规则。本规则是历史遗产,保留为参考文档,不再是执行入口。
影响:本规则 alwaysApply: false,实际已通过 role-menu Rule 14 → session-bootstrap 执行。 如需修改经验感知行为,修改 session-bootstrap.mdc B1/B2/B4,不再修改本文件。
⚠️ 本规则不再 alwaysApply(P0-3 修复,2026-03-20) 目的:减少每次对话的 token 注入(原 ~3000 token/次),行为不变。 role-menu.mdc 强制规则第14条保证:任何实质性任务完成后,AI 必须执行本规则的三步收尾。
⚠️ 同步维护说明:本文件的 Step 1/1.5/2/3 在
.cursor/rules/role-menu.mdc强制规则第14条中也有内联副本(因 hook 是 alwaysApply:false,role-menu 是 alwaysApply:true,两处需保持一致)。修改本文件任意 Step 时,必须同步修改 role-menu.mdc Rule 14 对应描述。
经验自动感知钩子(auto-experience-hook)
核心原则:执行时专注执行,结束时回顾沉淀。 不要在任务中途监控信号——结束后回顾比中途打断更可靠。
触发信号定义(A-E 类)
以下信号类型用于「任务结束回顾」步骤中的判断标准:
信号 A:踩坑信号
→ 遇到了报错/失败,且找到了解决方案(不是放弃)
信号 B:新发现信号
→ 完成了某个步骤,但这个步骤在现有 Skill 流程中没有覆盖
→ 判断标准:如果下次不记得,同类任务会重蹈覆辙
→ 特别触发:若本次对话发现 `_内部总控/知识库/` 目录下有文件,视为信号B
(该目录已废弃,正确位置:`认知结构/L1_系统性文档/[维度]/知识库/`;参见 `_内部总控/知识库放置规范.md`)
信号 C:步骤偏差信号
→ 实际执行与 Skill 描述的步骤顺序/内容不一致
信号 D:触发词不准信号
→ 某个 Skill 本该触发却没触发,或触发了不该触发的 Skill
信号 E:缺失 Skill 信号
→ 遇到某类任务,没有对应 Skill,AI 凭经验执行
→ 且这类任务未来可能重复出现
信号 G:结构性根因信号(元反思层)
→ AI 发现本次失败/偏差的根因不是执行疏忽,而是规范/机制本身设计有缺陷
→ 判断标准(满足任一即触发):
· 某行为今天又出错了,但类似情况之前已经出现过(≥2次同类失败)
· 发现某条规范只在某个 Skill 里,但这条规范应该普遍生效(应是 alwaysApply Rule)
· 发现某个重要文档/规范,AI 没有通过 D0/Rule 获知,靠「记忆」才知道要读
· 发现规范写的是「建议/应该」,但实际需要「禁止/必须」级别的强制
· 发现某条规范的覆盖范围写错了(太窄/太宽),或表述有歧义
→ 记录格式(追加到 PENDING-EXPERIENCES,类型写「结构性根因」):
| [今日日期] | [触发信号的 Skill 名] | 结构性根因 | [缺陷类型: 连接缺口/类型错误/执行力度/覆盖错误/表述歧义]——[一句话描述:如「role-Skill产品经理没有D0行,不读原则文档,是连接缺口而非执行疏忽」] | 🔲 待处理 |
→ ⚠️ 注意:触发信号G时,PENDING-EXPERIENCES 的记录格式要含「缺陷类型」字段,方便 project-retrospective 识别为 E6 类条目
信号 F:用户认知输入信号
→ 用户在对话中表达了具有学习/演进价值的思考、原则或洞见
→ 判断标准(满足任一即触发):
· 用户使用「应该...」「肯定应该...」「这是个原则...」「这是个底层问题...」
· 用户使用「这里有个问题...」「我发现...」「根本原因是...」
· 用户对系统行为/设计模式做出规律性总结(如「所有X都应该Y」「AI总是Z」)
→ 判断排除:用户只是提问(「这是什么?」)或只是要求执行任务(「帮我写X」),不触发
→ 记录格式(追加到 PENDING-EXPERIENCES,类型写「认知碎片」):
| [今日日期] | [当前对话主题/任务名] | 认知碎片 | [用户核心洞见一句话] | 🔲 待处理 |
→ 后续处理:project-retrospective 遇到「认知碎片」类条目时,调用 cognitive-capture-fragment 写入 L2,候选为 L1.5 原则
→ 注意:本次对话若已直接写入了 L2 碎片,PENDING-EXPERIENCES 可标注「已写入L2,见F-XXX」
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.
- 8d ago First seen · 421 lines · 121 tokens per session scan A a613793085f6
auto-experience-hook is a cursor rule published in the GitHub repository TashanGKD/tashan-cursor-skills (20 stars, last pushed 5mo ago), licensed MIT. It adds 121 tokens to every session and 7,472 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 cursor rules, from other repositories
10-feature-development
Feature implementation workflow and engineering mindset.
11-template-conventions
Reusable building blocks shipped with this template - use them instead of writing new ones.
12-new-project
Workflow for starting a new app from this template - rebranding, identity, cleanup and first feature.
13-updating-project
Workflow for updating an existing project - dependency and SDK upgrades, migrations, refactors, bug fixes.
01-tech-stack
Android tech stack and project standards.
02-architecture
Architecture rules - MVVM baseline, MVI for complex features.