AI Meeting Notes w/ Action Items + To-Do List Tracker

**Judgement:** This skill is a strong, opinionated solution for turning unstructured meeting content into structured markdown notes plus a simple, file-based to‑do system. It is particularly well-suited to users who already work with plain text/markdown and want a local-feeling alternative to SaaS meeting tools. **Key strengths:** - Clearly defined, rigid output format (title, summary, action items, decisions) improves consistency across meetings. - Auto-saving to `meeting-notes/` with normalized `YYYY-MM-DD_topic.md` filenames makes chronological organization and search straightforward. - Integrated `todo.md` workflow (add via "all" / "1,2,4", mark done, move, filter by owner, show overdue, daily "todo check") turns extracted items into a usable lightweight task system. - Works with many input styles (pasted notes, transcripts, VTT/SRT, chats, emails) without needing setup, accounts, or bots. **Main risks / limitations:** - The instructions are very strict (single-message response, exact filename pattern, specific template sections). LLM compliance may occasionally fail, leading to malformed filenames, missing prompts, or slightly off-format outputs. - Extraction quality (owners, deadlines, decisions) inherently depends on noisy input; the model may infer or miss tasks, mis-assign owners, or convert vague time references incorrectly. - Large or very long transcripts may stress context limits, forcing truncation or summarization that drops some details or action items. - The task numbering and sections in `todo.md` can drift if edits are made outside the skill or if multi-step operations are interrupted. **Recommended scenarios:** - Individual professionals, consultants, and small teams who want meeting minutes plus actionable follow-ups without subscribing to Otter/Fireflies-style services. - Teams that already store project documentation in git/markdown and want AI-assisted meeting capture that fits that workflow. - Users comfortable reviewing and spot-checking AI-generated action items, rather than treating them as authoritative records. **Less ideal for:** - Environments needing tight integration with external systems (calendars, ticketing, CRMs) or strict compliance/audit trails. - Users who prefer GUI-based task management rather than text/markdown files and command-like interactions (e.g., "done 3", "Sarah's tasks").