Are any agents always included in a scan?

Yes, the 'changelog-writer' is always included in partial scans to ensure that project history is consistently maintained regardless of the specific changes.

How are agents mapped to file changes?

The skill uses a predefined mapping table that connects file paths (like .github/ or plugins/core/) to specific agents such as delivery-lead, security-lead, or architecture-manager.

What happens if a Leader agent is triggered in partial mode?

When any Leader agent is triggered by a change, its corresponding Manager agent is also automatically triggered to provide the necessary strategic context.

How many agents are managed by this skill?

The skill manages a total of 15 agents: 3 Managers (strategic context), 10 Leaders (domain expertise), and 2 Writers (documentation output).

What is the difference between full and partial scan modes?

Full mode invokes all 15 agents across all tiers (Managers, Leaders, and Writers), while partial mode analyzes git diffs to select only the agents relevant to the specific files changed.

Agent Selection Manager

Name: Agent Selection Manager
Author: qmu

byqmu

0•

학습 및 문서화

Automates the selection of specialized documentation agents based on git diff analysis and scan modes.

This skill optimizes the Workaholic workflow by intelligently determining which specialized agents—Managers, Leaders, and Writers—are required for a repository scan. By analyzing code changes via git diff, it allows for efficient 'partial' scans that target only relevant domain experts, while supporting 'full' scans for comprehensive project analysis. It maps specific directory changes to corresponding roles such as the security-lead, infra-lead, or ux-lead, ensuring that documentation and strategic context are always synchronized with the latest codebase state.

주요 기능

010 GitHub stars

02Intelligent agent tiering across Managers, Leaders, and Writers

03Automatic dependency mapping between Leaders and Managers

04Domain-specific agent targeting for security, UX, and infrastructure

05Git-diff based partial scanning to reduce processing overhead

06Standardized JSON output for integration with automated workflows

사용 사례

01Automating domain-specific reviews when sensitive directories like security or database configs are modified

02Optimizing CI/CD pipelines by running only relevant agents on targeted pull requests

03Generating comprehensive documentation audits using the full 15-agent suite

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add qmu/workaholic select-scan-agents

For use in Claude.ai and ChatGPT

Download Skill