What problem does it solve? Saved bookmarks, watch-later lists, and saved notes pile up faster than anyone can review them, and titles or snippets rarely justify what deserves real attention. This Skill turns a bounded collection of saved items into a ranked, read-only triage queue with honest evidence labels, so you know what to read, skim, defer, or discard without mutating the source collection. ## Core Features & Use Cases - Conservative Deduplication: Normalizes URLs and stable IDs while preserving every original locator, keeping uncertain duplicates separate and flagged rather than silently merged. - Evidence-Labeled Classification: Assigns each item one of discard, skim, study, act, defer, or archive, with an evidence-coverage label (full content, snippet, metadata, user context, or judgment) and disclosed confidence. - Time-Budget Queue Selection: When given a time budget, selects a queue whose disclosed estimated total fits, moving worthwhile overflow to defer instead of ranking everything immediate. - Use Case: You have 80 saved videos and articles from the past month and only two free hours. Run the triage with your interests and time budget to get a ranked queue of what is actually worth watching, what to defer, and which links are dead or duplicates. ## Quick Start Triage my saved bookmarks from the last month against my current projects with a two-hour time budget and show me the ranked queue with evidence labels.
