home/writing/multi-agent-prompt-system-spec

MULTI AGENT PROMPT SYSTEM SPEC

GPTClaudeGemini··648 copies·updated 2026-07-14
multi-agent-prompt-system-spec.prompt
# Multi-Agent Prompt System — Restructure Spec
**Date:** 2026-03-19
**Author:** Tisone
**Purpose:** Define the full modular prompt system for all agents in the BMAD workflow
**Status:** Draft

---

## Problem Statement

All agent prompts are currently monolithic single files. An LLM cannot reliably
hold role identity, workflow rules, gate checks, and session state simultaneously
in one file. Under pressure or pushback, rules decay. Agents collapse roles.
Executors self-certify. Advisors become Executors. The Coordinator has no clear
reference for their own responsibilities.

The solution is a modular prompt system — each concern in its own file, each file
independently useful, all files organized into agent-specific modules.

---

## System Structure

when to use it

Community prompt sourced from the open-source GitHub repo TisoneK/bmad-crew (no explicit license). A "MULTI AGENT PROMPT SYSTEM SPEC" style prompt — adapt the placeholders and specifics to your task. Imported as-is and not independently retested here, so check the output before relying on it.

tags

writingcommunitygeneral

source

TisoneK/bmad-crew · no explicit license