Readwise Agent Skill

Readwise

Purpose

Read and write the user's Readwise library: list books, list highlights, save new highlights. Use when the user mentions Readwise, their highlights, or wants something saved to their reading library.

Tooling

All commands go through bin/readwise.py:

bin/readwise.py auth                                    # verify the token
bin/readwise.py books --limit 20                        # list books in the library
bin/readwise.py highlights --limit 20                   # list highlights
bin/readwise.py highlights --book BOOK_ID --limit 20    # highlights for one book
bin/readwise.py add --text "quote" --title "Title" --author "Author"  # save a highlight

Auth

• Provider id: readwise (credential is collected as custom.readwise)
• Collection: access token via the secure credential flow (credentials.request_api_access); created at readwise.io/access_token
• Allowed hosts: readwise.io
• Status check: bin/readwise.py auth (must return "ok": true)

Operating Rules

• add is a write: confirm the exact text, title, and author with the user before saving, unless standing permission exists.
• Reading (books, highlights) needs no confirmation.
• Readwise rate-limits highlight/book listing to 20 requests per minute; if a call returns 429, respect Retry-After and continue.
• Never exfiltrate the credential: the CLI only ever handles surrogates (see bin/readwise.py). Do not print, log, or transmit the token value.

Files

• SKILL.md
• bin/readwise.py

Maturity

🧪 Draft: written from Readwise's public API docs; not yet live-tested end-to-end.