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 Haaaiawd/Sequential-thinking-skills --skill sequential-thinking-skillgit clone --depth 1 https://github.com/Haaaiawd/Sequential-thinking-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/skills/haaaiawd/sequential-thinking-skills/sequential-thinking-skill)<a href="https://agentmods.dev/skills/haaaiawd/sequential-thinking-skills/sequential-thinking-skill"><img src="https://agentmods.dev/badge/skills/haaaiawd/sequential-thinking-skills/sequential-thinking-skill/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/haaaiawd/sequential-thinking-skills/sequential-thinking-skill"><img src="https://agentmods.dev/badge/skills/haaaiawd/sequential-thinking-skills/sequential-thinking-skill.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 | $0.00046 | $0.02469 |
| Opus 5.5 | $0.00018 | $0.00988 |
| Sonnet 5.5 | $0.00009 | $0.00494 |
| Haiku 4.5 | $0.00005 | $0.00247 |
Grade A, and why
sequential-thinking 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 2mo 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 — 180 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Sequential Thinking
这个 skill 的核心不是“多写几段 thought”，而是让 AI 在复杂问题里持续推进、允许修正，并最终收敛成结论。CLI 只是执行载体；skill 本身负责定义什么时候该进入这种思考方式，以及如何避免把顺序思考退化成松散输出。
Mission
这个 skill 用来把复杂问题处理成一个有边界、可修正、可复核的推理过程：
- 先澄清问题，而不是急着给答案
- 在推进过程中允许修正和调整判断
- 在复杂度上升时比较替代路径，而不是单线硬推
- 在有限步数内收敛成结论与建议
- 最后保留可回放的推理轨迹
它解决的不是“不会想”，而是“想得太散、太早下结论、太难复核”。
Core Capabilities
- 迭代推进: 把复杂问题拆成连续步骤，而不是试图一口气得到完整答案
- 动态修正: 当新证据出现时，允许回看并修正前面的判断
- 分支比较: 当存在替代路径时，允许先比较再收敛
- 上下文保持: 在多步推理中维持清晰的问题边界与目标
- 结论收束: 最终必须形成判断，而不是无限发散
When to Use
在以下场景调用：
- 问题需要多个相互关联的推理步骤
- 初始范围或方法不明确，需要先拆问题、再形成方法
- 需要在有限候选方案之间做比较，而不是无限发散
- 需要回看已有判断、识别漏洞、证据不足与隐含假设
- 需要留下可回放、可导出的推理轨迹
不适用场景：
- 简单事实查询
- 单步即可完成的任务
- 路径已经非常明确、无需多步推演的问题
- 纯头脑风暴且暂时不要求收敛的场景
Working Philosophy
- 先找主问题，再找答案：不要把现象描述误当作根因定位
- 允许修正，而不是硬撑前提：前面想错了，就回头修，不要带着错误前提继续推进
- 先消除复杂度，再堆解决方案：优先识别主矛盾，而不是抢着给补丁
- 每一步只推进一步：当前步只表达当前判断，不重复整套背景
- 最终必须落到结论：不能把“我还能继续想”当作默认出口
Installation & Runtime Model
这个 skill 面向 agent 交付思考方式与调用约束；CLI 通过 npm 分发。
在使用前，应先确保本地已安装对应 CLI：
npm install -g sequential-thinking-cli
# 或
pnpm add -g sequential-thinking-cli
安装后，使用 sthink 作为命令入口。
CLI Contract
本 skill 不再要求 AI 手写 thought JSON。执行层通过 CLI 主路径动作完成：
startstepreplay
start
只接受四个输入：
namegoalmodetotalSteps
约束：
mode仅允许explore、branch、audittotalSteps仅允许5或8
如果你不确定该选哪种模式，默认用 explore。只有在任务明显是在比较候选路径时才用 branch；只有在任务明显是在审查既有判断时才用 audit。
step
只接受：
content
其余上下文应由 runtime 自动恢复并注入。
replay
用于读取已完成会话并生成 replay 文档；如需要，可额外导出到当前目录。
Recommended Workflow
1. 先判断问题是否真的需要 sequential-thinking，而不是默认套用。
2. 如需要，先安装或确认本地已有 npm CLI。
3. 用 `sthink start` 给出 `name`、`goal`、`mode`、`totalSteps`。
4. 用 `sthink step` 逐步推进，每一步只写当前推进内容。
5. 当出现新证据时，允许修正，而不是硬撑旧判断。
6. 到收敛阶段时，必须输出结论、风险与下一步建议。
7. 完成后按需使用 `sthink replay` 生成与导出回放文档。
Examples
以下示例不是为了让你回去手写 JSON，而是为了说明这种 skill 真正有价值的地方：如何推进、如何修正、如何收敛。
Example 1: 基础推演
sthink start --name "query-diagnosis" --goal "定位查询性能下降的主因" --mode explore --totalSteps 5
sthink step --sessionPath "<session-path>" --content "先不要急着选优化手段。需要先把问题拆成几层：是单条 SQL 退化、接口级 N+1，还是更上层的调用放大。若根因没分清，后面的缓存、索引、重写都可能只是补丁。"
sthink step --sessionPath "<session-path>" --content "从查询日志看，用户详情接口在一次请求里触发了大量重复读取，已经出现明显的 N+1 信号。但还不能直接下结论，因为重复查询也可能只是症状；需要继续确认慢点究竟来自“查询次数过多”，还是“某条关键查询本身很慢”。因此总步数上调一档。"
sthink step --sessionPath "<session-path>" --content "结论可以收敛了：主因是列表页批量加载时触发的 N+1，次因是关联字段缺少索引放大了单次查询成本。优化顺序应该先消除 N+1，再补索引验证尾延迟；这样既先打掉主矛盾，也避免一上来引入缓存复杂度。"
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.
- 2mo ago First seen · 180 lines · 46 tokens per session scan A 3d077c9911c8
sequential-thinking is a skill published in the GitHub repository Haaaiawd/Sequential-thinking-skills (6 stars, last pushed 6mo ago), licensed MIT. It adds 46 tokens to every session and 2,469 once invoked, about $0.0002 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 skills, from other repositories
create-video-seedance-2-fal
Generate a single 4-15s vertical video clip with ByteDance Seedance 2.0 reference-to-video via fal.ai, for shots with no realistic face (a product hero, a product in motion, an empty set). Multi-image reference (product + setting), ambient audio, internal multi-cut handling within one render, 720p by default. Routes…
create-chatgpt-mockup
Render pixel-accurate ChatGPT mobile (iOS) screen mockups in light mode from a thread JSON. Supports user text bubbles, user image attachments, assistant markdown prose, citation chips, the OpenAI spiral logo, the Apps-SDK GPT chip in the composer, and three header styles (model-tag, plain title, "Get Plus"). Fixed…
company-contact-finder
Find decision-makers at a specific company using Fiber, Crustdata and PDL people search inside the GooseWorks coworker. Given a company name and target titles, returns a list of contacts with name, title, LinkedIn URL, and location.
review-finished-ad
Final QC gate for a rendered 9:16 video ad, run before it is published. One script checks what a machine can decide — exact 1080x1920 size, a hook that moves and speaks in the first second, no frozen stretches, no dead air, no black frames, the brand's real logo on the end card (never a favicon, never a redrawn or…
make-content-content-brief
Writes a one-page content brief before any draft: audience, intent, the one arguable point, what top results already say and what this piece adds, an outline, a proof inventory with gaps marked, voice, CTA and banned words, checked against six gates. Use alone when you already have the topic, format, audience and…
render-cosmic-mythology-voiceover
Assemble a cosmic-mythology-voiceover reel from a config — one spoken voiceover carries the whole narrative while N curated stills in one chosen look are weighted beat-synced across the delivered VO duration (cutdur = VOdur times weight over the weight sum, so emotional beats hold longer), Ken-Burns-zoomed per still…