Miniflux

**Judgement:** A solid, focused skill for agents and users who already have (or can provision) a Miniflux server. It offers rich, agent-friendly control over RSS/Atom feeds—especially strong for search, triage, and long-article handling—while relying on Miniflux for sync and storage. **Strengths:** - Good functional coverage: list/search articles, read content (with pagination), mark read/unread, inspect feeds/categories, show stats, and refresh feeds. - Very agent-friendly CLI design with clear flags, JSON output, brief/summary/full modes, and explicit pagination for large articles. - Config via env vars or config file is well-suited for automated environments. **Key Risks / Limitations:** - Requires an existing Miniflux instance and valid API key; not a generic RSS fetcher on its own. - Destructive operations (mark-read/mark-unread) can change state; poorly designed agents could mis-mark items if prompts are imprecise. - Network/API errors and Miniflux downtime will directly impact reliability; the skill surfaces errors but cannot mitigate backend issues. **Recommended Scenarios:** - Users or teams with Miniflux who want an AI/agent to manage and summarize their reading queue. - Research assistants that search across subscribed feeds, pull relevant articles (possibly in chunks), and generate summaries or notes. - Automation flows that periodically export unread items as JSON for classification, tagging, or ingestion into other knowledge systems, then mark them as read once processed.