Automated summary
Builds the daily ML/AI paper and news picks for the PaperTrace feed and updates the site's TypeScript content files.
Strengths
- Enforces a strict anti-hallucination rule that restricts every number to arXiv abstracts or official sources
- Specifies exact field formats (why, whyZh, tags, newsItems) that drop straight into daily.ts and page.tsx
- Provides prioritized sources (HuggingFace Papers, arXiv, official blogs) plus a validation checklist with npm run build
Limitations
- Tightly coupled to one project's layout and types (PaperTrace: src/lib/daily.ts, src/app/daily/page.tsx)
- Assumes the model can browse the web to verify official sources, which is not always available
- Does not automate retrieval: selecting papers, extracting numbers, and translating remain manual steps
Use when you need to add today's verified AI paper and news picks to the PaperTrace daily page.
Avoid on unrelated projects, or when official sources are unreachable and numbers would have to be cited without verification.
Security analysis
CautionThe skill uses Bash and network access for legitimate purposes: fetching official sources, editing two content files, running a build, and committing the digest. There is no destructive command, secret exfiltration, obfuscated payload, or safety-disabling instruction. Because it grants broad command and network capabilities, it is best rated caution rather than safe.
- •Declares broad Bash access; workflow only describes using it for `npm run build` and `git commit`.
- •Uses WebFetch/WebSearch against external sites, then edits repository source files based on fetched content.
- •No explicit instruction to avoid unrelated file edits or to scope changes strictly to the listed files.
Examples
Update the PaperTrace daily page with today's digest: pick 2–3 recent ML papers from HuggingFace Papers and arXiv, add them to src/lib/daily.ts, and run npm run build to verify.Add a news entry about the latest OpenAI model announcement to the newsItems array in src/app/daily/page.tsx, using only facts and numbers from the official OpenAI blog post, then verify with npm run build.Review the newest entries in PaperTrace's daily feed and flag any why/description text that uses 'reportedly', '~', or '据报' so I can replace them with verifiable numbers from the abstracts.name: daily-digest description: Add today's ML/AI paper and news picks to the PaperTrace daily feed. Use when the user wants to update the daily page with new content. allowed-tools: Read Edit Bash Glob WebFetch WebSearch
PaperTrace · 每日日报
为 PaperTrace 生成当天的日报内容,更新到 src/lib/daily.ts 和 src/app/daily/page.tsx。
⛔ 反幻觉原则(所有内容均适用)
论文条目(dailyPapers)的数字必须来自:
- arXiv abstract 页面(作者、标题、arXiv ID)
- 论文 abstract 本身(主要结果数字)
新闻条目(newsItems)的内容必须来自:
- 官方公告(官方博客、官方 HuggingFace 页面、官方 X/Twitter 账号)
- 官方技术报告或 README
绝对禁止:
- 用第三方媒体报道的数字填写 why/description(例如 "据报~81%")
- 把推测性技术细节当作事实(例如训练硬件)
- 用 "reportedly" / "据报" 等模糊措辞掩盖无法验证的来源
如果官方数字暂时无法获取,宁可省略该数字,不写模糊断言。
1. 找内容
Paper 来源(按优先级):
https://huggingface.co/papers ← 社区人气榜,首选
https://arxiv.org/list/cs.LG/recent
https://arxiv.org/list/cs.CL/recent
https://paperswithcode.com/latest
新闻来源(仅限官方):
官方 HuggingFace 模型页 README
官方博客(openai.com, anthropic.com, deepmind.google, ai.meta.com 等)
官方 X/Twitter 账号(@openai, @anthropic, @deepseek_ai 等)
官方技术报告 PDF
筛选标准(论文):
- 解决清晰问题,不是 survey
- 有明确数字结果(从 abstract 中可直接提取)
- 与受众相关:LLM / Diffusion / RL / Efficient Training / Reasoning
目标:每日 2–5 篇 paper + 0–2 条 news,宁缺毋滥。
2. 生成论文字段
拿到 abstract 后,只使用 abstract 中明确出现的数字:
论文标题:[title]
arXiv ID:[XXXX.XXXXX]
Abstract:[粘贴]
请输出 JSON:
{
"titleZh": "中文标题",
"tags": ["从以下选 ≤3 个: Diffusion LM, Pre-training, Fine-tuning, LoRA, Efficient Inference, Reasoning, Alignment, RLHF, Multimodal, Theory, Benchmark, Code, Agent, RAG, Quantization, MoE, Vision"],
"why": "英文,≤25词,格式:[做了什么] — [为什么重要],数字必须来自 abstract",
"whyZh": "中文,≤30字,同格式"
}
3. 新闻条目格式
新闻加到 src/app/daily/page.tsx 的 newsItems 数组顶部:
{
date: "YYYY-MM-DD", // 官方公告日期
title: "英文标题", // 准确描述事件
titleZh: "中文标题",
description: "...", // 仅含官方来源确认的事实
descriptionZh: "...",
source: "Official Blog / HuggingFace / ...", // 明确标注来源
sourceUrl: "https://...", // 直接链接到官方来源
tag: "Release" | "Research" | "Industry",
}
Description 写法原则:
- 用官方公告原文改写,不添加推测
- 数字直接引用官方数据(例如从 HuggingFace README)
- 如确实不确定某个数字,直接省略,不写 "~" 或 "reportedly"
4. 论文条目格式
加到 src/lib/daily.ts 的 dailyPapers 数组顶部:
{
date: "YYYY-MM-DD",
title: "英文原标题",
titleZh: "中文标题",
authors: "第一作者 et al.",
arxivId: "XXXX.XXXXX", // 只填 ID;非 arXiv 论文用 paperUrl 字段
paperUrl: "...", // 非 arXiv 时填(如 HuggingFace PDF 链接)
tags: ["Tag1", "Tag2"],
why: "...", // 数字必须来自 abstract 或官方 README
whyZh: "...",
pick: true, // 最多 1 篇
slug: "paper-slug", // 如果已有精读页则填
}
5. 验证
npm run build
- 确认新条目在
/daily顶部 - 中英文切换正常
- 检查 why/description 中无 "reportedly" / "~数字" / "据报" 等模糊措辞
快速清单
[ ] 确认内容来源为官方(论文/公告/README)
[ ] 从 abstract 提取数字,不从第三方报道引用
[ ] 加 paper 条目到 daily.ts 顶部
[ ] 加 news 条目到 page.tsx newsItems 顶部(如有)
[ ] npm run build 验证
[ ] git commit: "feat(daily): YYYY-MM-DD digest"
From the file to a first result
The download contains exactly the content displayed on this page. This copy may differ from the latest version in the source repository.
View the source and supporting filesThis download contains SKILL.md only. Any scripts, references, templates and dependencies are not included. Get the required files from the source, preserving their directory structure, before continuing.
1. Download the file
Download SKILL.mdFrom the root of a test project, save the file as .skill-downloads/daily-digest-mrwakdmu/SKILL.md, or use the command below. This folder is for inspecting the skill before installation.
Command for macOS/Linux or WSL: requires a POSIX shell and curl. On Windows without WSL, use the download button and create the folder manually.
mkdir -p '.skill-downloads/daily-digest-mrwakdmu' && curl --fail --show-error --location 'https://skills-guides.com/api/skills/daily-digest-mrwakdmu/download' --output '.skill-downloads/daily-digest-mrwakdmu/SKILL.md.part' && mv '.skill-downloads/daily-digest-mrwakdmu/SKILL.md.part' '.skill-downloads/daily-digest-mrwakdmu/SKILL.md'2. Check prerequisites
Read SKILL.md before using it. If it references scripts or external services, inspect them and prepare the required access in your test environment.
Tools declared by the author (check availability in your agent): Read Edit Bash Glob WebFetch WebSearch
No recognized prerequisites section in this file. This does not mean it is self-contained: libraries, tools and access requirements may be mentioned elsewhere in the content or source repository.
The platforms on this page are catalog indications. The paths below follow tool documentation; they do not prove this skill has been run in those tools. No minimum tool version is confirmed for this listing.
3. Load the skill
After inspection, copy the complete folder into your project using one of the commands below, then open a new tool session. The command stops if a folder with the same name already exists: compare the files before replacing it.
Claude Code
test -f '.skill-downloads/daily-digest-mrwakdmu/SKILL.md' && test -s '.skill-downloads/daily-digest-mrwakdmu/SKILL.md' && mkdir -p '.claude/skills' && test ! -e '.claude/skills/daily-digest' && cp -R '.skill-downloads/daily-digest-mrwakdmu' '.claude/skills/daily-digest'OpenSkills
OpenSkills must already be installed. The local install/read procedure is checked with OpenSkills 1.5.0, which requires Node.js ≥ 20.6.0. Catalog slugs are not resolved automatically.
Official OpenSkills installationopenskills install './.skill-downloads/daily-digest-mrwakdmu'openskills read 'daily-digest-mrwakdmu'Reading should display the installed file content and its base directory. It does not run the skill or install its dependencies.
For an agent that reads AGENTS.md, then sync the index. This command changes AGENTS.md: review the diff before using it.
openskills syncManual method or another tool: attach SKILL.md to a conversation, or open it in an agent that can read files. Include any required supporting resources. No universal folder is recognized by every tool.
4. Try a first task
Use a short, synthetic input: an API endpoint, a small document or an interface brief. In a new conversation, attach the file or give the agent its path, then adapt this request:
Read .skill-downloads/daily-digest-mrwakdmu/SKILL.md and apply its instructions to this example: [describe a test input]. Identify any missing files, dependencies or access. Produce the result requested by the skill without inventing data or running unreviewed scripts.Expected result: output matching the skill instructions, or a precise list of missing prerequisites. A read confirmation alone does not validate a first use. If an external resource is unavailable, that path remains untested.
Writing Fragments
Content
Mine raw writing fragments through a relentless interview, without imposing structure. Capture ideas, quotes, and leading words for later use.
UI Copywriting & Auditing
Content
Write or audit user-facing UI copy including buttons, errors, empty states, and tooltips across the monorepo, following the design system source of truth.
Music MIDI Extraction
Content
Extracts the MIDI or score from an existing music production via the CogMusic MCP tool and confirms availability.