Octop is a self-hosted, multi-user AI assistant that runs multiple specialized agents and connects them to chat interfaces, tools, and external services. It is for individuals, families, and teams who want a locally operated assistant with shared experts and persistent capabilities. Catalogue add-ons extend its agent and assistant workflows.
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 TencentCloud/Octop --skill guideline-section-expansiongit clone --depth 1 https://github.com/TencentCloud/OctopWrote 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/tencentcloud/octop/guideline-section-expansion)<a href="https://agentmods.dev/skills/tencentcloud/octop/guideline-section-expansion"><img src="https://agentmods.dev/badge/skills/tencentcloud/octop/guideline-section-expansion/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/tencentcloud/octop/guideline-section-expansion"><img src="https://agentmods.dev/badge/skills/tencentcloud/octop/guideline-section-expansion.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 | $0.00082 | $0.03126 |
| Opus 5 | $0.00041 | $0.01563 |
| Sonnet 5 | $0.00016 | $0.00625 |
| Haiku 4.5 | $0.00008 | $0.00313 |
Grade A, and why
guideline-section-expansion 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 4d 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 — 108 lines — stays where its author put it; the contents beside it link to each section on GitHub.
指南章节展开与学习路径图
用户反馈里最常见的两句话是「指南查到了但展不开」和「能不能把指南总结成流程」。本 skill 处理这两件事,并把它们严格限定为学习产物。
最小必读集
- 用户只要求总结筛查、随访、质控等某一专题时,本 skill 已内嵌学习边界、常用信源规则和输出字段;不再读取
source-policy.yaml、compliance-boundary.md或output-templates.md。最终仍由validate_output.py使用完整策略校验。 - 用户要求完整章节展开或学习路径图时,再按需读取
../../references/compliance-boundary.md和../../references/output-templates.md。 - 上述均为确定路径;不得通过
glob、grep或递归ls定位。
先分清用户要哪一种
| 用户说法 | 使用 |
|---|---|
| 展开这一章、细看、详解、这节讲了什么、看全文 | 章节展开 |
| 总结指南中某一专题的学习要点,例如筛查、随访、质控 | 章节展开 |
| 怎么学这份指南、学习顺序、总结成流程、给我路径 | 学习路径图 |
| 我掌握得怎么样、测一下、薄弱点 | guideline-learning-diagnosis skill |
| 建立每日连续学习 | guideline-learning skill |
命中「总结某一专题的学习要点」时,直接跳到「专题总结快速路径」;该节已给出完整执行规则,不再另外读取参考文件。
用户说「流程」时必须先确认他要的是学习顺序。如果他实际想要的是「遇到这种病人先做什么再做什么」,那是床旁处置请求,按 compliance-boundary.md 的高风险口径拒绝,并说明可以改为学习路径图。
前置:必须先有可核验的原文
展开和路径图都必须绑定已核验的权威文件:完整名称、发布机构、版本/年份、权威原文链接。
- 取不到权威原文时,不得凭记忆展开。只能说明「未取得可核验权威原文」,并给出候选来源方向。
- 完整章节或路径图只能使用 source-policy.yaml 白名单内的最终依据;专题总结按下方快速路径内嵌的同等规则执行。聚合平台默认只作线索;仅在原始正文明确访问受限、同时取得 S/A 正式元数据并完成标题/机构/版本/DOI/全文完整性/现行状态核验时,source-policy.yaml 指定的 B+ 平台才可作为正文承载渠道。
- 用户明确指定国际指南,或国内现行正式文件未覆盖主题时,读取
../../references/international-guideline-source-routes.yaml。普通补充只从 WHO、NICE、ESC、ADA、KDIGO 默认池选择最匹配的 1 站;明确涉及 GINA、GOLD、ACG、EASL 等体系时直接访问对应官网。首站确实不覆盖时最多增加 1 站,不得并行遍历。不得用国际建议静默覆盖中国大陆现行规范。 - 版本必须写清。指南改版频繁,展开旧版时要标注这是哪一版。
章节展开怎么做
完整章节展开使用 output-templates.md 的「指南章节展开」模板;专题总结直接使用下方快速路径的字段清单。要点:
- 给出原文定位:章号、节号、条目号或页码范围,让用户能回到原文核对。这是本能力的核心价值。
- 按原文结构讲,不重排成自己的逻辑;说明本节在指南整体结构中的位置和与上下游章节的关系。
- 只做结构化摘要,不整段复制全文。目标是帮用户读懂,不是替代原文,也要避免版权问题。
- 标注证据等级、推荐强度(原文有则写,没有不编)。
- 点明易混淆处——这是医生真实的学习痛点。
- 结尾必须声明不替代原文,以权威原文为准。
允许展开指南中的定义、诊断标准、鉴别框架、分层概念、检查与随访要求、质控指标、章节间的概念关系。不得把这些条款代入某个真实患者得出结论,不得改写成院内可执行文件。
专题总结快速路径(性能硬约束)
当用户要求面向基层总结筛查、随访、质控等单一专题时:
- 默认只选 1 份能够直接覆盖问题的现行主指南。用户未要求跨指南比较时,不主动扩展老年专项、基层专项或其他并行文件;只有主指南确实缺少该主题时才补第 2 份。搜索、抓取、B+ 降级、草稿校验和改稿共享同一份累计工具预算,任何阶段都不得重新计数。
- 直接读取
../../references/verified-source-entrypoints.yaml,先按aliases匹配主题。命中时直接web_fetch该条目的canonical_url,页面标题和 DOI 均匹配后才可作最终依据;不得用本地表项替代原文核对。 - 命中条目且当前日期未超过
recheck_after时,该次原文抓取即是实时核验,不再广泛搜索。只有用户明确询问「最新/现行/替代」,或已超过recheck_after,才再加 1 次限定官方域名的时效检索。 最终来源行必须使用命中条目的canonical_url;official_index_url只辅助核对卷期/DOI,不取代原文链接。订阅墙导致只能核对摘要或公开附录时,必须把未能核对的具体条款标为「待核验」,不得用二手内容补全。 2.1 未命中高频入口、但主题能对应专业学会或专科分会时,直接读取../../references/professional-society-source-routes.yaml,按分会名称或aliases选择一个最相关的官方域名做第 1 次限定检索;不得并行遍历全部分会。标记为official_discovery_pending_fulltext_acceptance的域名只用于找到完整正式原文,不能直接作为最终来源。 - 原始正文遇到 412、验证码、JS 阻断、超时或正文不可获取时,才允许 B+ 受控降级:先用原始发布方/正式期刊元数据页核对标题、机构、年份/版本和 DOI/文号,再抓取指定 B+ 平台的完整正式正文并检查完整性与现行状态。最终同时列“原始元数据”和“正文承载”两条来源;只有 B+ 页面或任一字段不一致时立即停止,不生成学习要点。
- 未命中入口表时,正常目标为 1 次定向搜索 + 1 次正文抓取,整个用户回合的总工具预算硬上限为最多 3 次
searchfree_search和 3 次web_fetch,包括失败、超时、空结果、国际补充、B+ 降级、C 级全网转述、校验后补链和改稿。每层最多一次定向查询:国内 A → 国际 A(仅在用户指定或国内无覆盖时)→ B 级发现/B+ 承载或 C 级全网。不得换旧版文件或猜测 URL 继续试探。找到 1 份能覆盖主题的现行正式主指南后立即停止搜索。 - 预算内未取得现行权威原文时,不得生成“原文展开”或声称掌握原文。若 C 级网页明确给出原始权威期刊/机构、完整标题、年份/版本及 DOI/文号信息,可输出单独的“背景转述”失败回执,并按
source-verify的 C 级格式披露;否则答复「未取得可核验的现行权威原文,本次不生成学习要点」。C 级转述不得涉及精确推荐、证据等级或药品高风险信息。 - 最终依据默认 1 份、最多 2 份:1 份现行主指南;仅在主指南确实不覆盖问题或用户明确要求比较时补 1 份专项指南。只接受正式指南、共识、规范或质控文件;综述、原始研究、科普、转载、会议材料和搜索摘要不得作最终依据。
- 常用国内原始来源域名包括
.gov.cn、.nhc.gov.cn、.ndcpa.gov.cn、.chinacdc.cn、.cma.org.cn、.medjournals.cn、.cmda.net、.cpma.org.cn、.pmph.com、cs.medjournals.cn、bi-dmp-cs.yiigle.com和rs.yiigle.com。国际 A 级域名以international-guideline-source-routes.yaml为准。搜索结果只用于定位原文;起草前先选定原文详情页或正式附件链接。通用列表页只能作候选,不能在无法对应具体文件时当作原文链接。 - 未命中入口表时,第 1 次查询必须同时包含主题、「最新/现行」和注册表匹配到的专业分会/学会域名;没有专业路由时才优先
site:rs.yiigle.com与site:cs.medjournals.cn。第 2 次查主管部门或正式期刊平台;第 3 次只用于确认版本替代关系或在原始正文访问受限后定位 B+ 完整正文。必须先确认现行版本,再摘要内容;新版已发布时不得回退到旧版凑答案。 - 草稿中所有 URL(包括正文和来源行)必须属于原始来源白名单,或满足“原始元数据 + B+ 正文承载”的成对降级规则;不得先写入 CDN、搜索结果或其他非白名单 URL,再依赖校验器反复修改。
- 输出校验因辅助来源域名或文档类型失败时,删除该辅助来源并使用已经核验的主指南重新起草;累计预算已耗尽时不得继续搜索补链。直接按以下字段生成一份完整草稿:【指南章节展开|主题】、依据、章节、原文定位、原文要点、章节关系、易混淆点、学习提示、边界、来源。不先写半成品,不为猜模板再读其他文件。
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.
- 4d ago Changed · +9 lines c92d781f0d5c
- 11d ago First seen · 99 lines · 82 tokens per session scan A cc17605ebce0
guideline-section-expansion is a skill published in the GitHub repository TencentCloud/Octop (1,533 stars, last pushed today), licensed MIT. It adds 82 tokens to every session and 3,126 once invoked, about $0.0004 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
manim-composer
Trigger when: (1) User wants to create an educational/explainer video, (2) User has a vague concept they want visualized, (3) User mentions "3b1b style" or "explain like 3Blue1Brown", (4) User wants to plan a Manim video or animation sequence, (5) User asks to "compose" or "plan" a math/science visualization.…
documentary-montage-production
Provider-independent workflow for AI agents producing documentary-style montages, archive-driven timelines, interview-supported sequences, nonprofit or advocacy shorts, historical explainers, and hybrid generated/archive edits. Use for editorial thesis development, research/source logs, fact-checking, archive rights…
explainer-video-production
Provider-independent explainer video production for AI agents creating educational, product, concept, nonprofit/public-service, onboarding, training, animated, mixed-media, and short social explainers. Use when an agent must turn a topic, brief, research corpus, product, process, policy, dataset, or complex idea into…
educational-animation-production
Provider-independent production workflow for AI agents creating educational animated lessons, classroom explainers, STEM visualizations, history/social-science animations, training modules, microlearning clips, whiteboard-style explainers, diagram-driven videos, and generated or assembled instructional animation…
manim-explainer-animation
Provider-independent production workflow for creating Manim-based explainer animations. Use when an agent must plan, code, render, QA, or hand off precise math, science, data, diagram, algorithm, or concept animations with Manim, including storyboard, scene architecture, formulas, coordinate systems, voiceover timing…
yaml
Skill "yaml" from dhaupin/vant, covering yaml, when to use, what to do, 1. syntax and 2. anchors.