skills.rest
mylukin
mylukinCOMMUNITY
Explore all skills by this author →
WebsiteGitHub

foreman-spec

Coordinate four AI agents to translate requirements into structured task breakdowns.

247|20|Updated Nov 28, 2025
Search Tags:#multi-agent#ai-orchestration#web-research#task-breakdown#requirements-analysis#mode-quick-deep#foreman-spec
One-click install
npx skills add https://github.com/mylukin/agent-foreman --skill foreman-spec
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: foreman-spec
Source: https://github.com/mylukin/agent-foreman/tree/main/plugins/agent-foreman/skills/foreman-spec
Command: npx skills add https://github.com/mylukin/agent-foreman --skill foreman-spec

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

This Skill orchestrates four specialized AI agents (PM, UX, Tech, QA) to transform high-level requirements into a structured, executable task breakdown. Each agent conducts web research before analysis to ground recommendations in industry best practices, case studies, and current trends, dramatically reducing planning time and architectural ambiguity.

Core Features & Use Cases

  • Research-first multi-agent analysis: Each agent performs web research prior to analysis to ground decisions in real-world data.
  • Four specialized roles: agent-foreman:pm (Product Manager), agent-foreman:ux (UX/UI Designer), agent-foreman:tech (Technical Architect), agent-foreman:qa (QA Manager).
  • Mode options: Quick Mode (parallel, ~3 minutes) and Deep Mode (serial, ~8-10 minutes) with Q&A after each agent to refine outputs.
  • Phase-driven workflow: Includes Phase 0: Mode Selection and Phase 1: Codebase Scan to adapt analysis to the project state.

Quick Start

Use the foreman-spec skill to break down a high-level requirement into a 4-agent task and research plan. Example: "Create a feature to improve onboarding conversions" → generates four agent prompts (PM, UX, Tech, QA) with web research and a stepwise task breakdown. Start in Quick Mode for fastest results, with a combined final Q&A.

Frequently Asked Questions about foreman-spec

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I break down high-level requirements into concrete tasks using multiple AI agents?▼

Task breakdown with multiple AI agents involves coordinating specialized roles—Product Manager, UX Designer, Technical Architect, and QA Manager—each conducting web research before analysis. Foreman-spec orchestrates these four agents in parallel (Quick Mode, ~3 minutes) or serial (Deep Mode, ~8-10 minutes) to transform requirements into structured, actionable task specifications grounded in industry best practices.

What's the difference between Quick Mode and Deep Mode for requirements analysis?▼

Quick Mode runs all four agents in parallel with a combined final Q&A, completing in ~3 minutes for rapid task breakdowns. Deep Mode executes agents serially with Q&A refinement after each agent, taking ~8-10 minutes but producing deeper analysis. Choose Quick Mode for speed; Deep Mode when architectural or design decisions need thorough vetting.

Can AI agents analyze my existing codebase when generating task specifications?▼

Yes. Foreman-spec includes Phase 1: Codebase Scan using Glob pattern matching to adapt the four-agent analysis to your project's current state. This enables each agent—PM, UX, Tech, and QA—to ground their task recommendations in your existing architecture, patterns, and dependencies rather than generic guidance.

Do the AI agents use web research when analyzing requirements?▼

Yes. Each of the four specialized agents conducts web research before analysis to ground recommendations in current industry best practices, case studies, and trends. This research-first approach dramatically reduces planning ambiguity and ensures task specifications align with real-world solutions.

What output does foreman-spec produce from a high-level requirement?▼

Foreman-spec generates a structured task breakdown consisting of four agent-specific prompts (PM, UX, Tech, QA), each with web research findings and stepwise task recommendations. The result is an executable specification that spans product strategy, UX design, technical architecture, and QA verification—all indexed to your selected mode and codebase context.

Is prior architecture or product design knowledge required to use multi-agent task breakdown?▼

No prerequisite expertise is required. The four specialized agents—Product Manager, UX Designer, Technical Architect, and QA Manager—each bring domain expertise and conduct independent web research. You provide the high-level requirement; the agents handle domain-specific analysis, research, and task generation across all four perspectives.

Related Skills

Explore More Product & Management Skills→
nechmadsnechmads

workflow-planning-01-create-prd-interactive

Generate a PRD from guided discovery interviews and save it to .agents/product/PRD.md.

Community
Basic
xiaoweidotnetxiaoweidotnet
22

planning-and-task-breakdown

Decompose software requirements into executable tasks with acceptance criteria and dependency mapping.

Community
Advanced
FaisalAlqarniFaisalAlqarni
5

superpowers-research-mode

Explore codebases with Grep and Glob for pattern identification.

Community
Basic
bumfobumfo
1

deep-research

Orchestrate parallel web, codebase, and third-party code research with GoT-driven tasks.

Community
Advanced