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 gy910210/hermes-research-skills --skill paper-reader-zhgit clone --depth 1 https://github.com/gy910210/hermes-research-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/gy910210/hermes-research-skills/paper-reader-zh)<a href="https://agentmods.dev/skills/gy910210/hermes-research-skills/paper-reader-zh"><img src="https://agentmods.dev/badge/skills/gy910210/hermes-research-skills/paper-reader-zh/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/gy910210/hermes-research-skills/paper-reader-zh"><img src="https://agentmods.dev/badge/skills/gy910210/hermes-research-skills/paper-reader-zh.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 | $0.00080 | $0.01063 |
| Opus 5 | $0.00040 | $0.00531 |
| Sonnet 5 | $0.00016 | $0.00213 |
| Haiku 4.5 | $0.00008 | $0.00106 |
Grade A, and why
paper-reader-zh 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 12d 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 — 88 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Hermes 适配说明
- 本 skill 现面向 Hermes 使用,优先依赖 Hermes 原生工具:
search_files、read_file、write_file、patch、session_search、delegate_task、cronjob、browser、web/search、vision。 - 若正文提到
references/...或scripts/...,优先读取当前 skill 目录下对应文件,不再依赖 Claude 专属目录结构。 - 原始 Claude
agents/openai.yaml不作为执行前提;需要并行研究、分工精读或角色评审时,改用 Hermes 的delegate_task。 - 保留原有研究方法论与产物契约,但执行层统一按 Hermes 工具体系落地。
Paper Reader(中文)
输入
sources[]: 论文、官方博客、benchmark 页面、dataset page、PDFdepth:"concise"|"full"(默认concise)
输出
paper_notes- 字段契约见
../research-orchestrator-zh/references/artifact-contracts.md
- 字段契约见
步骤
- 明确来源类型:论文、官方博客、benchmark 页面、dataset page 或其他。
- 记录
publication_time、venue_or_source、institution_or_company;无法确认时标明未披露。 - 抽取问题设定、任务类型、方法/系统结构、关键数据集与 benchmark。
- 只记录来源明确写出的结论与结果,整理为
explicit_claims[]与results[]。 - 对论文记录训练设置、评测协议、消融与失败案例;对官方页面记录产品能力边界与未披露项。
- 补充
reusable_for_sections[],说明这些内容最适合回写到哪些综述章节。 - 如果论文明显属于“方法/训练样本流密集型”论文,额外抽取:
raw_data_objectstoken_objectsstagewise_trainingpublic_prompt_or_template_evidencedisclosed_vs_undisclosed_ratios
守护
- 明确区分“来源显式陈述”和“阅读者推断”。
- 引用要带页码、章节、段落或 URL 线索。
- 对 benchmark、release 状态、数据规模、是否开源等硬事实,优先保守。
- 对 prompt/template、样本比例、过滤阈值、sampling 规则尤其保守;没看到正文/附录原文就不要写满。
Hermes 强执行协议
调用这个 skill 时要真正完成“精读记录”:
- 先判断来源类型:本地 PDF、本地 Markdown、网页、官方博客、benchmark 页面、dataset page。
- 本地文件优先:
- Markdown / txt 用
read_file - PDF 优先走 OCR/文档提取相关能力,再进入精读
- Markdown / txt 用
- 网页优先使用
web/search+ 抽取;页面结构复杂时再用browser。 - 输出不能只是一段摘要,必须整理成结构化
paper_notes。 - 每条核心结论都尽量带出处线索:页码、章节名、表格号、URL、小节标题。
- 若用户给了多篇来源且内容较长,使用
delegate_task分篇精读,再在主线程合并结果。
推荐的 paper_notes 结构
建议至少包含:
source_typetitlepublication_timevenue_or_sourceinstitution_or_companyproblem_statementtask_typemethod_or_systemdatasets_and_benchmarksexplicit_claimsresultslimitations_or_undisclosed_itemsreusable_for_sections
若是方法/训练样本流密集型论文,再补:
raw_data_objectstoken_objectsstagewise_trainingpublic_prompt_or_template_evidencedisclosed_vs_undisclosed_ratios
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.
- 12d ago First seen · 88 lines · 80 tokens per session scan A f01b081acede
paper-reader-zh is a skill published in the GitHub repository gy910210/hermes-research-skills (2 stars, last pushed 4mo ago), licensed MIT. It adds 80 tokens to every session and 1,063 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 2026-08-31.
Other skills, from other repositories
moai-docs-generation
Documentation generation patterns for technical specs, API docs, user guides, and knowledge bases using real tools like Sphinx, MkDocs, TypeDoc, and Nextra. Use when creating docs from code, building doc sites, or automating documentation workflows.
graphic-ebook
Creates professionally designed B2B SaaS e-books in HTML + CSS, exported as print-ready PDF. 3–10 pages, 9 style presets, 11 page layout types. Trigger when user says "create an ebook", "design a lead magnet", "make a PDF guide", "build a gated content piece", "write a B2B ebook", "design a white paper", "create a…
Read, create, merge, split, fill, watermark, stamp, secure, inspect, render, and visually verify PDF files, including AcroForm workflows.
ocr-and-documents
Extract and structure text from scanned PDFs and document images with a local-first OCR workflow, layout-aware fallbacks, and explicit heavyweight dependency controls.
PDF Processing Pro
A PDF-processing workflow for handling forms, extracting tables, and using OCR to turn scanned text into searchable data. It also performs automated data checks using built-in scripts.
law-to-markdown
A legal-document conversion skill for turning PDF, Word, or plain-text regulations into structured Markdown. Markdown is a text format that uses simple symbols for headings and lists.