autonomy-type-based
由 luciusrockwing 提供
Search & Research简介
Type-based autonomous task queue system.
描述
该技能的功能
A task queue system that routes agent work by task type. The agent autonomously handles research, writing, and analysis tasks pulled from a tagged queue file, while maintenance, backup, and security tasks are left to cron jobs. Tasks carry priority labels that determine selection order within each type.
Prevents the agent from spending tokens on routine maintenance that scheduled cron jobs can handle more cheaply and reliably.
何时使用
- Queuing competitor research tasks for the agent to work through during the day
- Auto-drafting blog posts and documentation from a backlog without manual prompting
- Separating daily log cleanup and backups from AI-driven content work
- Running weekly metrics analysis on a heartbeat without human triggering
- Managing a prioritized writing and research backlog across multiple daily sessions
示例工作流
User asks: 'I added a competitor pricing research task to QUEUE.md with @priority:high @type:research'
- 1
Read tasks/QUEUE.md and find all items in the Ready section
- 2
Filter tasks to only those tagged @type:research, @type:writing, or @type:analysis
- 3
Select the highest-priority qualifying task (competitor pricing research)
- 4
Mark task as In Progress in QUEUE.md and begin research
- 5
Save findings to tasks/outputs/ and log a summary to memory/[today].md, then move task to Done Today
Structured research report with findings, sources, and recommendations saved to tasks/outputs/competitor-pricing-analysis.md
安装
clawhub install luciusrockwing/autonomy-type-based运行 CLI 命令,或把这段提示词粘贴到 Claude Code、Cursor、Codex 或你在用的任意编程智能体里。