todoist-api

This skill provides procedural guidance for reading and managing Todoist data through the `td` command-line tool. It solves the problem of turning natural-language task-management requests into precise CLI invocations, covering full CRUD operations across tasks, projects, sections, labels, and comments. The skill emphasizes machine-readable output for agents (`--json`, `--ndjson`, `--full` flags) so results can be parsed and acted on programmatically.

The documentation walks through quick-add natural-language parsing (due dates, priorities, projects, sections, labels), listing and filtering tasks, viewing task details, creating tasks with explicit flags, and updating, completing, or reopening tasks. A dedicated reference file documents Todoist's filter query language, including date, priority, label, project, section, and assignment filters plus logical operators for combining them. A companion reference covers completed-task retrieval. A key safety feature is the explicit confirmation requirement: before any destructive action (deleting, completing, updating, or archiving), the skill instructs the agent to ask the user for confirmation, while read-only operations proceed freely.

The target users are people who manage their to-do lists in Todoist and want an AI assistant to query, add, or reorganize tasks conversationally. Typical use cases include adding tasks from natural language, surfacing overdue or high-priority work, filtering tasks by project or label, and bulk-reviewing or completing items. It assumes the `td` CLI is installed via npm and authenticated over OAuth.