ChosenXu

Newsletter Link Harvester — Development skill for Claude Code

Development community

Agent Skill: harvest links from newsletter emails into Raindrop.io with the author editorial context attached - Gmail read-only, three-layer dedup, zero-context library compare.

How to install Newsletter Link Harvester

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

What Newsletter Link Harvester does

Agent Skill: harvest links from newsletter emails into Raindrop.io with the author editorial context attached - Gmail read-only, three-layer dedup, zero-context library compare.

Alternatives in Development

Gmail Sync — /gmail-sync - Sync Application Status from Gmail 36.6k ★
Mono Color Skill — One-ink editorial print image skill — warm paper, halftone photography, active negative space, and restrained 3k ★
Gmail AutoAuth MCP Installation Guide — This guide will help you install and configure the Gmail AutoAuth MCP server for managing Gmail operations thr 1.1k ★

README

Newsletter Link Harvester

English | [简体中文](README.zh-CN.md)

An [Agent Skills](https://agentskills.io)-compatible skill that **harvests links from newsletter emails and files them into a [Raindrop.io](https://raindrop.io/) library — with the author's own introduction attached to every link**. Works with Claude Code, Codex CLI, Gemini CLI, GitHub Copilot, Cursor, and WorkBuddy.

You read a design/tech newsletter once a week; the good links die inside your mailbox. This skill turns each issue into a permanent, searchable bookmark set: every link lands in a per-sender sub-collection, its note carries `via: · ` plus a one-sentence summary of *why the author recommended it*.

What it does

Filter — searches Gmail by a sender whitelist (primary) and optional subject keywords (secondary), within a look-back window (default 7 days). Explicitly naming an older issue extends the window on demand.
Parse — two proven channels: Markdown direct links (Quail-style) and plain-text redirect wrappers (SubStack-style [ url ] tokens, resolved to their final destinations before anything is saved).
Context — extracts each link's editorial context: the author's introduction or recommendation, taken from the sentence (or adjacent paragraph) that contains the link. Saved links never lose their "why".
Dedup (3 layers) — in-batch (URL normalization: www / utm / trailing slash / fragment / case), cross-run (a local state file of processed email IDs), and against your whole Raindrop library.
Confirm → Save — a full preview with target sub-collection naming; nothing is written without your explicit approval. Bookmarks are batch-created with via: + context notes. Promotional entries (self-promotion, social channels, unsubscribe) are flagged and excluded by default.

Highlights

Zero-context library compare — the whole Raindrop library is paged straight into a local file over HTTP (9 requests for a 1,300-bookmark library); b