Task Management

Overview

spark task covers the full task lifecycle: create, implement, dispatch, sync. It can drive an AI agent (the kimi CLI) to auto-implement issues, and dispatches tasks into dedicated working directories for development.

Core Capabilities

Task Initialization

Create the standard task directory structure (tasks/issues/) inside a project.

spark task init

Issue File Management

Create, list, and delete issue descriptor files. Issue files are Markdown-format task descriptions.

spark task create my-feature                     # Create an issue
spark task create my-feature --content "description"    # With content
spark task list                                  # List
spark task delete my-feature                     # Delete

AI Implementation

Automatically implement the task described in an issue via the kimi CLI.

spark task impl my-feature
spark task impl my-feature --tui    # Interactive mode

Task Dispatch and Sync

Dispatch a task to a dedicated working directory; sync the implementation back when development is complete.

# Dispatch to a working directory (auto-init Git, create GitHub repo)
spark task dispatch my-feature

# Sync back to the task directory once development is done
spark task sync my-feature

Workflow

create → impl → dispatch → (develop) → sync
  ↑                                      ↓
  └────── tasks/issues/*.md ←───────────┘

Parameters

Dependencies

• git and the gh CLI (dispatch needs the GitHub API)
• The kimi CLI (required for impl)

Related

• Task Command Spec
• Task Usage Guide