td-task-management

To maintain AI agent task state across context window resets, you need persistent task tracking that logs progress, decisions, and blockers externally. This ensures work resumes exactly where it left off when the session ends or context resets.

Structured handoff between AI sessions is the process of transferring task state, logs, and decisions from one agent to another. It works by saving progress externally so the next agent picks up with full context.

To start tracking work for a new AI agent session, initialize the task manager and check the current status. You can start a new session and view the current task status by running the `td usage --new-session` command.

Yes, you need the 'td' CLI binary installed and available in your system's PATH to manage tasks across ephemeral context windows. This dependency is required for logging progress and performing structured handoffs.

Yes, you can manage multiple issues in a single AI work session. The workflow management supports single-issue focus, multi-issue work sessions, and code review processes to handle various development scenarios.

You should use CLI task management for AI agents when work spans multiple context windows or requires handoffs. It prevents context loss during complex features by automatically logging decisions and blockers that manual notes often miss.