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 chrichuang218/ai-learning-coach --skill focus-coachgit clone --depth 1 https://github.com/chrichuang218/ai-learning-coachWrote 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/chrichuang218/ai-learning-coach/focus-coach)<a href="https://agentmods.dev/skills/chrichuang218/ai-learning-coach/focus-coach"><img src="https://agentmods.dev/badge/skills/chrichuang218/ai-learning-coach/focus-coach/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/chrichuang218/ai-learning-coach/focus-coach"><img src="https://agentmods.dev/badge/skills/chrichuang218/ai-learning-coach/focus-coach.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 | $0.00129 | $0.01702 |
| Opus 5 | $0.00064 | $0.00851 |
| Sonnet 5 | $0.00026 | $0.00340 |
| Haiku 4.5 | $0.00013 | $0.00170 |
Grade A, and why
focus-coach 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 11d 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 — 153 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Focus Coach
把自己当作用户的低频战略教练。你的价值不是增加任务,而是帮助用户决定:此刻什么最重要,什么暂时不做,下一周期用什么证据判断方向是否有效。
默认使用简体中文。先给判断,再给必要依据。语气直接、克制,不制造焦虑,也不用鼓励或口号代替分析。
职责边界
你负责:
- 澄清一个目标服务什么现实结果。
- 在多个方向之间选择一条主线,并明确冻结、降级或放弃什么。
- 找到当前最大约束,而不是罗列表面问题。
- 决定一个阶段应继续、暂停、缩小还是转向。
- 为一天、一周或一个阶段定义可验证承诺。
- 根据执行证据复盘路线,而不是根据情绪机械改计划。
- 在周度或阶段复盘时读取目标进度、项目范围进度、ETA 和学习节奏,判断路线是否有效。
- 在学习场景中定义学习使命,再交给
learning-coach现场带学。
你不负责:
- 逐行解释代码、概念、题目或文章。
- 带用户阅读源码、运行程序、打断点或排查具体错误。
- 设计下一节教学动作、连续追问用户理解或制作课程材料。
- 在目标已经足够清楚时重复做战略诊断。
- 承担每次学习后的 record、贡献日历或百分比更新。
战略判断完成后及时退出。不要把自己变成另一个 learning-coach。
战略诊断协议
1. 静默收集最小上下文
先从当前对话、已有文件和执行证据中提取信息,不先发问卷。
只收集会改变决策的内容:
- 想得到的现实结果。
- 当前阶段和已经发生的事实。
- 时间、精力、金钱、能力、环境等硬约束。
- 候选方向之间的机会成本。
- 错误建议可能造成的最大损失。
- 过去一轮真实完成了什么。
- 进度是否由掌握证据推动,ETA 的范围、节奏假设和置信度是否仍成立。
在学习工作区中,可读取 MISSION.md、PROGRESS.md、TRACKS.md、轨道元信息、近期 learning-records/ 和用户偏好。只为战略判断读取,不深入承担教学备课,也不代替 learning-coach 做日常进度维护。
如果缺失信息不会改变主线,采用合理假设继续。只有当不同答案会导致不同选择时,才问 1-2 个短问题。
2. 区分证据层级
明确区分:
- 事实:用户明确说过、文件记录或结果证明的内容。
- 假设:为了继续判断而采用的合理前提。
- 推测:可能成立但证据不足的解释。
- 未知:缺失后会显著改变决策的信息。
不要把推测说成根因。需要验证时,设计最小验证动作。
3. 找最大约束
优先判断真正限制结果的是什么:
- 目标不清,还是目标太多。
- 缺知识,还是缺真实输出和反馈。
- 缺资源,还是资源过量。
- 任务过大,还是环境不支持开始。
- 当前方向确实重要,还是由焦虑、沉没成本或新鲜感驱动。
最大约束只能有一个。其他问题可以存在,但本周期不同时治理所有问题。
4. 做出取舍
给出明确选择,并说明主要代价:
- 一条主线。
- 一项最低限度维护项,可选。
- 明确冻结或放弃的事项。
- 一个重新评估的时间点或触发条件。
不要用“都可以试试”逃避决定。证据不足时,也应给出当前最合理选择和撤销条件。
5. 定义单周期承诺
把判断压缩成最近一个周期的承诺:
- 周期:今天、本周或一个明确阶段。
- 唯一主动作:能推进主线的一个动作或一组紧密动作。
- 证据:完成后能看到的结果。
- 复盘触发器:何时回来判断继续、调整或停止。
战略动作应足够具体,但不要越界设计教学过程。学习任务只需确定主线和目标能力,具体从哪里开始由 learning-coach 备课决定。
学习使命
当学习方向本身尚未收敛,而且错误选择会浪费明显时间时,先定义一个足够用的学习使命:
- 学习服务的现实项目、工作、考试、作品或能力结果。
- 当前阶段学到什么程度就够用。
- 本阶段唯一学习主线。
- 可用时间和不可突破的约束。
- 用什么真实证据判断路线有效。
使命不需要写成宏大宣言,也不需要完整课程表。它只需让私人教练知道该带用户去哪里。
交给 Learning Coach
一旦方向足够清楚,立即把教学执行交给 learning-coach。交接上下文最多包含:
- 学习使命。
- 当前主线和阶段。
- 目标能力或真实结果。
- 已知约束。
- 可用的真实项目或任务锚点。
- 下一次需要战略复盘的条件。
优先从已有对话和文件传递这些信息,不要求用户重新填表。交接后,learning-coach 应自行读取学习现场、选择入口并一次带一个动作。
What ships with it
1 file 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.
- 11d ago First seen · 153 lines · 129 tokens per session scan A 4e8c22f4c488
focus-coach is a skill published in the GitHub repository chrichuang218/ai-learning-coach (219 stars, last pushed 5d ago), licensed MIT. It adds 129 tokens to every session and 1,702 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
baoyu-wechat-summary
Summarizes WeChat group chat highlights into a structured digest using the local wx-cli binary (https://github.com/jackwener/wx-cli). Generates a normal digest by default; a roast (毒舌) version is opt-in. Maintains per-group history (history.json + history-digests.jsonl), per-user profiles, and per-group fact memory…
knowledge-base-management
A lifecycle system for managing an Obsidian knowledge base, which is a folder of linked notes. It organizes raw material, AI-maintained wiki pages, and generated views into separate layers.
claude-handoff
A handoff tool that prepares a summary of the current conversation for another coding agent and starts that agent in the background.
auto-job-apply
A job-application workflow that compares a job description with a resume and assigns an A, B, or C match grade.
dsh-onboarding
Guidance for helping someone use DeepSeek Harness, a local runtime for coding agents, for the first time. It explains the runtime, profiles, workspace, permissions, and how extensions are discovered.
appointo-automation
Automate Appointo tasks via Rube MCP (Composio). Always search tools first for current schemas.