todo

# Todo Management Manage todo files in the `todo/` directory. Each todo is a markdown file named `{id}-{slug}.md` where `id` is a unique incrementing integer. ## Commands Based on $ARGUMENTS: - **create [description]** — Create a new todo file - **list** — List all current todos with their IDs and titles - **done [id]** — Mark a todo as completed by deleting the file - No arguments — List all todos (same as `list`) ## Determining the next ID To assign an ID to a new todo, you MUST check both: 1. Current files in `todo/` directory 2. Previously deleted files in git history: ```bash git log --all --diff-filter=D --name-only --pretty="" -- 'todo/*.md' | grep 'todo/' | sort -u ``` 3. Previously created files (in case they exist on other branches): ```bash git log --all --diff-filter=A --name-only --pretty="" -- 'todo/*.md' | grep 'todo/' | sort -u ``` The new ID must be **one higher than the highest ID ever used** across all of these sources. Never reuse a deleted ID. ### Known ID history IDs used so far (as of initial skill creation): - 1 (deleted), 2, 3 (deleted), 4 (deleted), 5-11 (never used — gap), 12, 12 (two files), 13, 14, 15, 16, 17, 18, 19, 20 Always verify by running the git commands above — do not rely solely on this list. ## File format Use this structure for new todos: ```markdown # {Short title} ## Context {How this was discovered — what real scenario or user workflow exposed the issue. Be specific: "While running tractor check on a 500-file repo..." not "There is a performance issue." This grounds the todo in real-world value and helps future readers judge priority.} ## Problem {What's wrong or what's needed — 1-3 paragraphs} ## Desired state {What done looks like — concrete and verifiable} ## Notes {Optional: implementation hints, related files, impact, priority} ``` The **Context** section is important — it answers "why should anyone care?" by showing the real situation where the limitation was hit. Without it, todos become abstract technical debt that's hard to prioritize. Keep it concise. The title in the filename slug should be a few hyphenated keywords (e.g., `21-query-diagnostics-empty-results.md`). ## Slug conventions - Lowercase, hyphenated - 3-6 words max - Descriptive enough to understand without opening the file ## Listing format When listing todos, show: ``` 2 streaming-mode 12 field-role-elements-all-languages 12 xmlnode-ir-remaining-work 13 semantic-transform-rewrite ... ``` Right-align the ID, then two spaces, then the slug (filename without ID prefix and `.md` extension).