How Agents Spawn
Task enters In Progress
The server builds a prompt from the task description, project context snippets, prompt template, and any file attachments.
CLI process starts
Claude Board runs
claude with --no-input and the configured permission mode. The working directory is set to the project’s configured path.Streaming begins
All output is captured and streamed to connected clients via WebSocket. Logs are persisted to the database.
Live Terminal
Each running agent has a live terminal you can open from the task card. It shows every action Claude takes — file reads, edits, bash commands, and reasoning. See Terminal for details.Token Tracking
Claude Board tracks token usage per task in real-time:| Metric | Description |
|---|---|
| Input tokens | Tokens sent to Claude (prompt + context) |
| Output tokens | Tokens generated by Claude |
| Cost estimate | Calculated from model pricing |
| Cache read/creation | Tokens served from or added to prompt cache |
task:usage WebSocket event and visible on the task card.
Smart Timer
The task timer measures active agent time:- Starts when the task enters In Progress
- Pauses when the task moves to Testing (awaiting review)
- Resumes if the task returns to In Progress after requesting changes
- Stops when the task moves to Done
The smart timer gives you an accurate picture of how long Claude actually spent coding, excluding review wait time.