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 agentmods add skills/julardepick/user-thoughts.skill/user-thoughtsnpx skills add JularDepick/user-thoughts.SKILL --skill user-thoughtsgit clone --depth 1 https://github.com/JularDepick/user-thoughts.SKILLWrote 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/julardepick/user-thoughts.skill/user-thoughts)<a href="https://agentmods.dev/skills/julardepick/user-thoughts.skill/user-thoughts"><img src="https://agentmods.dev/badge/skills/julardepick/user-thoughts.skill/user-thoughts.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 | $0.00058 | $0.05333 |
| Opus 5 | $0.00029 | $0.02667 |
| Sonnet 5 | $0.00012 | $0.01067 |
| Haiku 4.5 | $0.00006 | $0.00533 |
Grade A, and why
user-thoughts 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 5d 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 — 345 lines — stays where its author put it; the contents beside it link to each section on GitHub.
user-thoughts.SKILL
概述
跨会话、跨 Agent 时,用户积累的决策和约束会彻底丢失。user-thoughts.SKILL 将这些想法持久化到 mdbase,绑定于项目而非会话——任何接手的 Agent 读取 mdbase 即可继承用户的完整意图,无需重新推导。
重要声明
- 本 SKILL 系指本文档及同目录下内容,整体称
user-thoughts.SKILL - 用户:使用、调用、提及本 SKILL 的发言人
- 行为边界:SKILL 只做想法分析和记录,不干预 Agent 对用户指令的执行。这是核心设计——当用户说"把按钮改成红色"时,Agent 应同时执行修改并记录偏好,两者并行互不干扰。记录行为不应延迟或阻塞用户的实际工作
When to Use
- 用户输入
/user-thoughts或/ustht命令时 - 用户讨论架构、技术栈、规则、UI/UX 等项目决策时
- 用户发言中提及"想法"、"记录"、"mdbase"等关键词时
- 用户表达项目相关的需求、偏好、约束时
使用指南
- 用户发言包含项目想法、决策、需求、规则、偏好时,SKILL 自动激活
- 命令以
/user-thoughts或/ustht引导(详见 references/commands.md) - 自然语言触发兼容任何语言——Agent 应匹配用户意图,而非特定关键词
命令前缀:/user-thoughts 为完整前缀,/ustht 为简写,两者等价可互换。
激活边界:当用户发言同时包含指令和想法时(如"把这个按钮改成红色"),Agent 应执行指令并同时记录想法。仅当发言纯粹是闲聊或与项目完全无关时(如"今天天气不错")才不激活。
语言策略
- SKILL 本体(SKILL.md、references/、assets/)语言固定为中文,不随用户语言变化
- Agent 输出适配用户语言:命令反馈、sortin 摘要、mdbase 展示、提示信息等面向用户的输出,应使用用户当前对话所用的语言——用户无需切换语言即可理解反馈
- 想法原文保留:raw 记录和 mdbase 中的想法内容保持用户原始语言,不翻译、不转换——保持原文可避免翻译引入的歧义,且用户原始表述是最权威的意图表达
工作原理
工作流
用户发言 → Agent 识别项目想法 → 写入 raw/(即时计划)
↓
用户执行 /ustht sortin → Agent 处理 raw/ → 按维度追加到 mdbase/
↓
用户执行 /ustht mdbase show → Agent 展示组织好的想法库
工作模式
- 被动模式(
INSTANT_STATUS=off):仅响应技能命令,不自动识别想法 - 即时计划模式(
INSTANT_STATUS=on且SKILL_STATUS=on):自动识别用户想法并写入#raw/,mdbase 写入延迟到 sortin - 忽略模式:
ignore start/end区间内发言不记入 - 只读模式(必需工具缺失时):只读命令可用,写入命令返回提示
模式组合:即时计划 + 忽略可同时生效。SKILL_STATUS=off 时即时计划自动暂停,无论 INSTANT_STATUS 值如何。被动/即时计划通过 define.ini 持久化;忽略区间仅上下文有效。
目录定义
@/:SKILL 安装目录(SKILL.md 所在目录,即user-thoughts/)~/:工作目录#ustht/=~/.ustht/(.ustht是user-thoughts的运行时缩写)#mdbase/=~/.ustht/mdbase/#ignored/=~/.ustht/ignored/#raw/=~/.ustht/raw/#export/=~/.ustht/export/
#ustht 目录结构
.ustht/
├── define.ini # SKILL 运行状态
├── README.ai.md # .ustht/ 目录说明
├── raw/ # 用户原始发言(按日期分片)
│ └── yyyy-mm-dd.md
├── ignored/ # 被标记忽略的发言
│ └── yyyy-mm-dd.md
├── mdbase/ # 整理后的用户想法库
│ ├── backlog.md # 待办事项(用户明确提出但尚未开始的事项,由 sortin 归入)
│ ├── README.ai.md # mdbase 索引与概览
│ └── details/ # 按维度组织的想法文件
│ ├── rules.md # 项目规则、约束、原则
│ ├── plans.md # 项目规划、方向性想法
│ ├── ui/
│ │ ├── outline.md # UI 整体设计
│ │ └── details.md # UI 细节设计
│ ├── dev-stack.md # 技术栈选型、框架决策
│ ├── general.md # 通用(不属于其他维度的想法)
│ └── ... # 按需扩展
└── export/ # 从 mdbase 导出的内容
What ships with it
23 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.
- assets/Runtime-Template/define.ini 48 B
- assets/Runtime-Template/mdbase/backlog.md 380 B
- assets/Runtime-Template/mdbase/details/dev-stack.md 572 B
- assets/Runtime-Template/mdbase/details/general.md 176 B
- assets/Runtime-Template/mdbase/details/plans.md 429 B
- assets/Runtime-Template/mdbase/details/rules.md 453 B
- assets/Runtime-Template/mdbase/details/ui/details.md 490 B
- assets/Runtime-Template/mdbase/details/ui/outline.md 475 B
- assets/Runtime-Template/mdbase/README.ai.md 1.4 KB
- assets/Runtime-Template/README.ai.md 921 B
- references/commands.md 9.3 KB
- references/edge-cases.md 15 KB
- references/safety.md 4.7 KB
- references/sortin.md 8.2 KB
- scripts/common.py 3.1 KB runs code
- scripts/ignore_ops.py 4.8 KB runs code
- scripts/init.py 3.4 KB runs code
- scripts/show_mdbase.py 4.0 KB runs code
- scripts/show_raw.py 1.5 KB runs code
- scripts/sortin.py 13 KB runs code
- scripts/status.py 2.2 KB runs code
- scripts/toggle.py 2.7 KB runs code
- scripts/write_raw.py 3.9 KB runs code
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.
- 5d ago First seen · 345 lines · 58 tokens per session scan A 257d5296af99
user-thoughts is a skill published in the GitHub repository JularDepick/user-thoughts.SKILL (4 stars, last pushed 1mo ago), licensed MIT. It adds 58 tokens to every session and 5,333 once invoked, about $0.0003 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
archive
Persistent knowledge wiki and memory system that compounds across conversations using an Obsidian-compatible vault. Consult at the START of any non-trivial task to recall project context, preferences, and past discoveries. Ingest sources and build wiki pages when accumulating knowledge on a topic. Save at the END when…
project-wiki
组织和维护 Vibe Coding 项目的 wiki 文档体系。当用户要求初始化/重构 wiki 结构、进入新项目且 wiki/ 不存在、需要诊断和修复文档腐化、或需要管理 specs/refs/reviews 子目录的生命周期时触发。.
mindos-zh
MindOS 是用户的本地知识助手,也是跨会话、跨 Agent 共享的知识库。它保存决策记录、会议纪要、SOP、 排错经验、架构选型、调研结论和偏好设置。 仅 mindRoot 知识库内任务。不用于:改代码仓库、项目源码、KB 外路径。 核心概念:空间、指令(INSTRUCTION.md)、技能(SKILL.md);笔记可承载指令与技能。 触发场景:保存或记录任何内容、搜索历史笔记或上下文、更新或编辑文件、整理或重组文件结构、 执行SOP或工作流、捕获对话中的决策、复盘或总结经验、追加表格或CSV数据、跨Agent交接上下文、 提炼经验教训、同步关联文档、查找之前是否讨论过某事、查询历史决策、查找模板或SOP、…
mindos
MindOS: local knowledge assistant & shared KB. Keeps decisions, notes, SOPs, debugging lessons, research findings, preferences across sessions/agents. Core: save notes, search KB, organize files, run workflows, review, append CSV, hand off context, distill lessons. NOT for app source or paths outside KB. Triggers…
mindos-max
MindOS: local knowledge assistant & global memory layer. Keeps decisions, notes, SOPs, debugging lessons, architecture choices, research findings, preferences, conversation summaries for all connected agents. PROACTIVE: (1) search MindOS first for past context, (2) offer to save after valuable work, (3) persist key…
mindos
Operate a MindOS knowledge base: update notes, search, organize files, execute SOPs/workflows, retrospective, append CSV, cross-agent handoff, route unstructured input to the right files, distill experience, sync related docs. Use when the task targets files inside the user's MindOS KB (mindRoot). NOT for editing app…