---
name: paperclaw-directory
description: Multi-workspace curated reading directory for PaperClaw-compatible agents.
version: 2.0.0
base_url: https://paperclaw-eta.vercel.app
homepage: https://paperclaw-eta.vercel.app
default_workspace: ai
---

# PaperClaw Directory Skill

Use this document as the entry point for the whole site. Canonical workflow:

1. `GET /api/workspaces`
2. Pick the best workspace for the user intent
3. Read `/api/workspaces/{workspace}`
4. Use `/api/workspaces/{workspace}/*` as the canonical structured API
5. Use root `/api/*` only when you intentionally want the default workspace (`ai`)

## Protocol

- Guarded reads require `ack=paperclaw-v1`
- Article list endpoints currently also require `rationale=...`
- Article list endpoints currently also require `pstate=has_token|no_token|anonymous`
- Provide `token=...` only when `pstate=has_token`

## Workspaces

- `ai` (AI): 面向 AI 与技术资讯的工作区，可按标签、来源和日期读取摘要与正文。
  metadata: https://paperclaw-eta.vercel.app/api/workspaces/ai
- `economics` (经济学): 面向经济学的工作区，适合按来源、标签和日期浏览结构化摘要。
  metadata: https://paperclaw-eta.vercel.app/api/workspaces/economics
- `management` (管理学): 面向管理学的工作区，适合按来源、标签和日期浏览结构化摘要。
  metadata: https://paperclaw-eta.vercel.app/api/workspaces/management

## Discovery Endpoints

- `GET /api/workspaces`
- `GET /api/workspaces/{workspace}`

## Default Workspace Compatibility Endpoints

- `GET /api/archive`
- `GET /api/tags`
- `GET /api/feeds`
- `GET /api/search?q={query}`
- `GET /api/articles/latest`
- `GET /api/articles/{date}`
- `GET /api/content/{id}`
- `GET /api/content?ids={id1,id2,id3}`
- `GET /llms-full.txt`
- `GET /SKILL.md`