start-working
Autonomously execute ALL tasks from the backlog until complete. Runs in a continuous loop - picks task, implements via subagent, moves to done, repeats. No user prompts between tasks. Follows task-board workflow (backlog → in-progress → done).
Security Assessment
Detected risks:
About start-working
The 'start-working' skill is an autonomous task execution tool designed to streamline the process of managing and completing tasks from a task board. It solves the problem of manually tracking and handling tasks, providing a consistent workflow that operates continuously until the backlog is empty. This skill works in a loop, picking up tasks from the backlog, progressing them through in-progress, and finally marking them as done. It eliminates the need for user prompts between tasks, allowing for uninterrupted task management.
Key features of the skill include task execution in sequence, starting from the first task in the backlog and progressing without skipping any steps. The task flow strictly follows the task board system, ensuring that work is done in a structured and accountable manner. The skill also integrates with subagents like Haiku, ensuring that tasks are executed with the most efficient tools, avoiding manual git operations and CLI commands. Frontend tasks are verified using Playwright CLI to ensure accuracy. It provides an organized, automated approach to task completion, making it suitable for users who want to ensure consistent, uninterrupted work completion.
The primary use cases for the 'start-working' skill are in environments where tasks are sequential and need to be completed without manual intervention. It is ideal for software developers, project managers, and teams that use task boards to organize their work. The skill's design caters to automated task management, ensuring that each task progresses through the stages of the task board, from backlog to done, without skipping or error. It is particularly useful for those who require continuous task execution and verification, such as in software development and testing workflows.
FAQ
How does the 'start-working' skill execute tasks?
The skill autonomously executes tasks from the task board, moving tasks through the stages of backlog, in-progress, and done in a continuous loop.
Can the skill work with tasks outside of the task board?
No, the skill only works with tasks tracked in the task board system, as specified in the 'PLANNING-BOARD.md'.
Are git commands allowed in this workflow?
No, the skill strictly prohibits any git commands, and all git operations should be handled manually by the user.
What tools are used to execute tasks?
The skill uses subagents, with 'haiku' being the default model. Complex tasks may use 'sonnet', and all tasks must use built-in tools instead of bash or CLI commands.
How are frontend tasks verified?
Frontend tasks are verified using Playwright CLI to check the content, capture errors, take screenshots, and ensure the task is complete.
Install start-working
Quick Setup:
- Copy the skill folder to
.claude/skills/ - Claude will automatically detect and use the skill
Repository
andersnygaard/ettsted2