randyhaddad/briefing-loop

Briefing Loop is a plugin for AI coding agents (Claude Code, Codex, and others) that builds a personalized daily reading digest. You tell it which articles, newsletters, papers, podcasts, or YouTube videos to include, and it learns your preferences over time based on what you read, save, reject, or ask about. The selection logic is described as similar to how a music streaming service balances familiar recommendations with occasional surprises. Most of what it delivers matches your established taste, but a small portion is intentionally outside your usual areas to help it learn what else you might like. The digest can be delivered to several destinations: a chat interface, email, Slack, a PDF or EPUB file, or directly to a Kindle device for approved articles. The tool verifies sources, authors, publication dates, and recency before including items. Your preference profile and reading history are stored as files you can inspect and edit directly. Installation works through the Claude Code plugin marketplace using three commands, or as a Codex plugin via a manifest file in the repository, or as a generic skill folder for any agent harness that supports that format. After installation, you prompt your agent to set up the digest and provide it with links to start building your initial profile. Personal data such as delivery addresses, credentials, and reading history should not be committed to version control. The project is MIT licensed.