alonegg

Paperdaily Skills — Data skill for Claude Code

Data community

Claude Code skills & agent protocol for paperdaily — daily-paper KG platform.

How to install Paperdaily Skills

This entry records only its repository, not the path inside it, so there is no exact command to give. Open alonegg/paperdaily-skills and copy the folder into ~/.claude/skills/, or the file into ~/.claude/agents/.

What Paperdaily Skills does

Claude Code skills & agent protocol for paperdaily — daily-paper KG platform. Thin query CLI + deep-research pipeline (retrieve → OA full-text → agent deep-read → optional upload back to your workbench).

Alternatives in Data

Private GPT — Complete API layer for private AI applications on local models: RAG, skills, tools, MCP, text-to-sql, and more 57.5k ★
Azure Document Intelligence .net — Document text, table, and data extraction 1.8k ★
Claude Deep Research Skill — Enterprise-grade deep research skill for Claude Code with 8-phase pipeline, source credibility scoring, and au 1k ★

README

paperdaily-skills

[paperdaily](https://www.paperdaily.org) 是一个"重后端、轻前端"的每日论文系统：把论文流入知识图谱，按你的兴趣画像做每日 4 层推荐，并提供可编程的 v1 API。本仓库是它的公开 Claude Code skills 与 agent 协议文档。

**English.** This repository ships the public Claude Code skills and agent-protocol docs for [paperdaily](https://www.paperdaily.org), a daily research-paper recommendation platform with a knowledge-graph backend and a bearer-key v1 API. Two skills are included: `paperdaily` (thin query CLI: field digests, author lookups, semantic search over fuzzy research directions or seed papers, and watchlist management — queries are read-only, watchlist create/delete/run are writes and always ask you first) and `paperdaily-deep-research` (a three-stage pipeline that goes from a research field or a single seed paper to a citation-grounded literature review — platform retrieval, local open-access full-text fetching, and agent-team deep reading/synthesis, with optional upload of your own analysis back to the paperdaily workbench). All full-text fetching is OA-first, institutional access is strictly opt-in, downloaded PDFs never leave your machine, and nothing here touches Sci-Hub.

Skills 清单

目录	用途
`paperdaily/`	thin 查询 skill：领域/学科/话题日报、作者近作、语义检索（模糊研究方向 / 给定论文找相似）、watchlist 周追踪。查询只读；watchlist 的 create/delete/run 是写操作，调用前必征得你同意（需 `write:profile`，按需索取）。
`paperdaily-deep-research/`	深度调研 skill：从「一个领域 / 一篇种子论文」到「可溯源的深度文献综述」——平台检索推荐 → 本地 OA 全文瀑布下载 → agent team 逐篇精读 + 跨篇综合；可选把你自己的分析产物回传 paperdaily 工作台。
`mcp/`	MCP 服务（非 skill）：`paperdaily-mcp` stdio 服务器，把 v1 API 的 24 个端点暴露为标准 MCP 工具，任何兼容 MCP 的助手（Claude Desktop / Claude Code / Cursor …）都能直接接入。安装与配置见 `mcp/README.md`。

配套文档：

文件	内容
`AGENT_PROTOCOL.md`	agent ↔ paperdaily 服务端的对外契约（scopes、分层检索、回传、任务收件箱、行为规范、合规红线）。
`SKILL_SPEC.md`	Skill 编写十条规范（声明件、phase-gate、provenance、同意点、诚实账本……）与兼容矩阵。
[`docs/quickst