Does it require manual input between every task?

No, the skill is designed for full automation. It will continue looping through the task list until the entire specification is completed or a halt is triggered.

Can I monitor the tasks while they are running?

Yes, the orchestrator uses tmux to create dedicated windows for each sub-task, allowing you to attach to the session and view live execution logs.

Where is the task progress stored?

Progress is automatically saved to SEQUENTIAL_STATE.json and a human-readable SEQUENTIAL_PROGRESS.md file within your spec directory.

What happens if a task requires human intervention?

The system includes a HALT signal mechanism. If a task outputs a specific halt promise, the orchestrator pauses and waits for user input before continuing.

How does this skill prevent API quota exhaustion?

It executes tasks one at a time with configurable delays between iterations, ensuring your API usage stays within safe limits during long-running jobs.

Sequential Task Orchestrator

Name: Sequential Task Orchestrator
Author: PeterFile

byPeterFile

•

생산성 및 워크플로

Automates the serial execution of complex project specifications with real-time tmux visibility and intelligent rate-limit management.

The Sequential Orchestrator is designed for high-reliability execution of multi-step development tasks derived from Kiro specifications. It manages long-running jobs by processing sub-tasks one at a time, preventing API quota exhaustion and rate-limit triggers. By utilizing a Python-based loop and tmux sessions, it provides a transparent, windowed environment where developers can monitor progress or let the agent run autonomously overnight. The skill features automated state persistence, sub-agent task assignment for precision, and integrated safety halts for manual intervention.

주요 기능

01Real-time monitoring via organized tmux windows for each sub-task

02Persistent progress tracking via JSON and Markdown state files

03Configurable execution delays to avoid API rate limits and quota exhaustion

04Fully automated serial task execution without user intervention

0525 GitHub stars

06Intelligent task assignment using specialized LLM sub-agents

사용 사례

01Processing complex feature implementations from detailed Kiro specs

02Managing throttled sequences of agent tasks that require high stability

03Executing massive codebase refactors or migrations overnight

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add peterfile/devpilot-agents sequential-orchestration

For use in Claude.ai and ChatGPT

Download Skill