jc-english-chunk-cards
Skill1 starsChanged 19 days agoMITView on GitHub
把用户日常中文、英文对话、网页、截图或知识库过往文案拆成语义块,制成墨墨记忆卡并写入指定牌组。用于「丢一段工作英语给我」「写进记忆卡」「按 chunk 制卡」「从日常思考抽英语」「和外国人商务洽谈怎么说」「面试自我介绍英语」。空调用、只说「今天的」或「从聊天里抽」时,从当天 Grok 会话用户原文抽一个主题再制卡。从知识库抽句时只做一个主题,按把话说完需要的深度出连贯开口链,不套固定张数。不用于雅思/四六级词表、墨墨背单词云词本、知识库沉淀、改简历投递,或只问 Markji 制卡语法。
- Reads credentials
In this repository117 other files it gives its agents.
- vercel-react-best-practices/AGENTS.mdAGENTS.md
- xhs-visual-director-skill/AGENTS.mdAGENTS.md
- mattpocock-skills/CLAUDE.mdCLAUDE.md
- about-face-4Skill
- ab-test-setupSkill
- ad-creativeSkill
- ai-seoSkill
- analytics-trackingSkill
- b2b-sales-opportunity-followupSkill
- backend-concurrency-auditSkill
- ckm:banner-designSkill
- better-colorsSkill
- brand-guidelinesSkill
- ckm:brandSkill
- churn-preventionSkill
- claude-md-optimizerSkill
- clean-architectureSkill
- coach-sales-callSkill
- code-chain-study-commentsSkill
- cold-emailSkill
- competitor-alternativesSkill
- content-strategySkill
- copy-editingSkill
- copywritingSkill
- cross-project-delivery-acceptanceSkill
- daring-design-iterationSkill
- deploy-to-vercelSkill
- ckm:designSkill
- domain-skill-factorySkill
- domain-skill-loopSkill
- email-sequenceSkill
- form-croSkill
- free-tool-strategySkill
- git-worktreeSkill
- go-backend-clean-architectureSkill
- golang-gin-apiSkill
- golang-gin-databaseSkill
- gsap-coreSkill
- gsap-frameworksSkill
- gsap-performanceSkill
- gsap-pluginsSkill
- gsap-reactSkill
- gsap-scrolltriggerSkill
- gsap-timelineSkill
- gsap-utilsSkill
- guangzhi-b2b-salesSkill
- hatch-petSkill
- hueSkill
- impeccable-redesign-orchestratorSkill
- impeccableSkill
- ios-messaging-designSkill
- ip-as-logoSkill
- jc-audio-transcriberSkill
- jc-bookmarksSkill
- jc-client-howto-guideSkill
- jc-company-researchSkill
- jc-concept-deconstructSkill
- jc-daily-thought-social-publisher-v5Skill
- jc-decision-gradeSkill
- jc-design-teardownSkill
- The first 60 shown.
The file
Files
JC English Chunk Cards
把日常材料做成「正面整句英文、背面语义块」的记忆卡,写入墨墨记忆卡。不是背单词,也不是中译英默写。
首次回复
当前对话第一次触发时,读取 首次说明,按用户语言输出,然后立刻拆材料。后续不重复介绍。
材料已经是一句中文、英文、聊天或截图时,介绍完直接制卡并上传。用户没贴材料、只调用本 Skill、或说「今天的 / 从聊天里抽」时,走空调用:读取 从对话抽条,跑 harvest_today.py,再按 抽哪些句子 制卡。点名知识库、日常思考或过往文案时,读该材料并按同一份场景规范抽句。环境变量和 Skill 目录 .env 都没有 Token 时才问 MAIMEMO_TOKEN,获取地址见 记忆卡接口。不要先要词表、牌组清单或确认页。
边界
做:
- 从日常材料抽出完整英文句并切语义块;
- 按本 Skill 模板写成记忆卡;
- 写入用户自建牌组,默认名称
英语。
不做:
- 雅思 / 四六级 / BEC 词表 → 不制卡;
- 墨墨背单词云词本、
add_words、提交认识/模糊; - 知识库沉淀 →
jc-knowledge-capture; - 改简历、投递策略、JD 定制 → ChengziCV;用自我介绍和过往经历练开口英文仍走本 Skill;
- 只问 Markji 标签语法、不给学习材料。
工作流程
- 抽句
已贴原句:一句一卡。中文先译成美国人会说的句子。一段话拆成多张,不塞进一张,不套场景比例。
从知识库、日常思考、过往文案或对话抽条:读取 抽哪些句子。只做一个主题,按把话说完需要的深度出连贯开口链;用户指定场景则整批按该类。
无材料:先python3 scripts/harvest_today.py --days 1,按 从对话抽条 和 抽哪些句子 抽句,列出正面再上传。 - 切块与排版
读取 切块 和 卡片模板。正面纯英文;背面才分块。顺序从左到右往下加,不从句尾倒读,不拆主谓宾。 写入
读取 记忆卡接口。Token 先读环境变量MAIMEMO_TOKEN,没有再读本 Skill 的.env。不要把 Token 写进SKILL.md或 git。运行:python3 scripts/markji_upload.py --deck 英语 --file <cards.json>JSON 为
[{"content":"..."}]。grammar_version由脚本固定为 3。- 交付
报告主题、这条链覆盖哪些环节、张数、牌组名和每张正面。提醒在 App 里加入学习。失败则给出 HTTP 码和errors。
核心 Gotchas
- 建卡 body 不要带
deck/chapter,否则not a valid ObjectId。 grammar_version用 3;1 会报无效 GV。/markji/decks返回 403 是记忆卡开放能力未开通,不要改走背单词接口。- 正面写中文或
/分块,就测不到秒懂。 - 学习圈词必须给日常替换,否则用户会把 spaced review 说出口。
停止
材料是沉淀、改简历投递或考纲刷词时停止制卡并转交。Token 缺失只问这一项。牌组不存在时列出已有名称,不擅自新建(API 也不能建牌组)。
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

