一键导入
init
First-time project setup for Mighty Powers. Creates config file, artifact directories, offers to generate CLAUDE.md, and explains available workflows.
用 Codex 或 Claude 帮你安装 复制这段 Prompt,粘贴到 Codex、Claude 或其他助手里,让它检查 Skill 页面并帮你完成安装。
菜单
First-time project setup for Mighty Powers. Creates config file, artifact directories, offers to generate CLAUDE.md, and explains available workflows.
用 Codex 或 Claude 帮你安装 复制这段 Prompt,粘贴到 Codex、Claude 或其他助手里,让它检查 Skill 页面并帮你完成安装。
基于 SOC 职业分类
Grill the user relentlessly about a plan, decision, or idea. Use when the user wants to stress-test their thinking, or uses any 'grill' trigger phrases.
Use when you need to research a domain, market, or technical area before planning
Use when starting any conversation — establishes how to find and use skills, requiring Skill tool invocation before ANY response including clarifying questions
Audit frontend and backend implementation against every PRD file, one by one, using parallel subagents. Produces a full status table (done / partial / not-started), identifies critical-path blockers, and recommends next sprints. Use when you need to know where you stand against your PRDs.
Push the LLM to reconsider, refine, and improve its recent output. Use when user asks for deeper critique or mentions a known deeper critique method, e.g. socratic, first principles, pre-mortem, red team.
Produce the architecture: a lean spine of invariants that keeps everything built from it consistent, projected into whatever format the work needs. Use when the user says "create the architecture", "create technical architecture", "architecture spine", or "create a solution design".
| name | init |
| description | First-time project setup for Mighty Powers. Creates config file, artifact directories, offers to generate CLAUDE.md, and explains available workflows. |
Set up a project to use Mighty Powers effectively. Run this once when you first install the plugin in a project.
BMAD no longer ships a separate init skill — orientation is handled by help. Mighty Powers combines both: /mp:init scaffolds project config and directories; /help guides what to do next in the lifecycle.
/mp:init
Check what already exists:
.mighty-powers/config.yaml — project config (shared, commit this).mighty-powers/config.user.yaml — optional personal overrides (gitignore).mighty-powers/custom/ — optional skill customization TOML filesCLAUDE.md — project instructions for Claudedocs/plans/ — Quick Track plan directorydocs/planning/ and docs/implementation/ — Lifecycle artifact directories (or paths from config)package.json or equivalent — project metadataReport what was found and what's missing.
If .mighty-powers/config.yaml doesn't exist, create .mighty-powers/ first, then write:
# Mighty Powers project configuration
# See: https://github.com/anderson-0/mighty-powers
project_name: "<detected from package.json or directory name>"
user_name: "Developer"
communication_language: "English"
document_output_language: "English"
# Where planning artifacts are saved (PRDs, research, UX, architecture)
planning_artifacts: "docs/planning"
# Where implementation artifacts are saved (stories, sprint status, specs)
implementation_artifacts: "docs/implementation"
# Default output root for skills that reference {output_folder} (brainstorm-session, project-context, etc.)
# Usually matches planning_artifacts; override if you want a separate tree
output_folder: "docs/planning"
# Additional project knowledge files (read by lifecycle skills)
project_knowledge: "docs"
Ask the user to customize:
If config already exists but is missing output_folder, offer to add it (default to the same value as planning_artifacts).
Mention optional personal overrides: .mighty-powers/config.user.yaml for machine-local settings (gitignored).
After config exists (created or already present), read path values from it and ensure directories exist. Use defaults from Step 2 when keys are missing.
Create each path that does not exist yet:
mkdir -p .mighty-powers/custom
mkdir -p docs/plans
mkdir -p docs/planning
mkdir -p docs/implementation
Also mkdir -p for the resolved planning_artifacts, implementation_artifacts, and output_folder values when they differ from the defaults above.
Do not overwrite existing files — only create missing directories.
If no CLAUDE.md exists, offer to generate one:
"No CLAUDE.md found. Want me to analyze this project and create one?"
If yes, analyze the project:
package.json / Cargo.toml / go.mod / pyproject.toml for tech stackIf CLAUDE.md already exists, offer to add a mighty-powers section if it doesn't reference the plugin.
Check .gitignore for .mighty-powers/ handling. If missing or too broad, suggest:
# Mighty Powers local state (learnings, audit history, personal overrides)
.mighty-powers/*
!.mighty-powers/config.yaml
!.mighty-powers/custom/
Note: .mighty-powers/config.yaml SHOULD be committed (team project config). config.user.yaml, learnings, audit history, and other local state should stay gitignored.
Present a welcome message:
Mighty Powers initialized for <project_name>!
Created or verified:
.mighty-powers/config.yaml — project configuration
.mighty-powers/custom/ — optional skill overrides (team TOML)
docs/plans/ — Quick Track plans
docs/planning/ — lifecycle planning artifacts
docs/implementation/ — stories, sprint status, specs
CLAUDE.md — project instructions (if generated)
Not sure where to start?
/help — phase-aware guidance (BMAD-style "what should I do next?")
/mp:init — re-run setup if paths or config change
Quick Track (build features in this repo):
/brainstorm — design exploration with approval gates (mp:brainstorming)
/write-plan — wave-based implementation plan
/execute-plan — execute plan wave-by-wave
/quick-dev — fast flow for small changes
/investigate — systematic debugging
Lifecycle Track (formal planning from scratch):
/brainstorm-session — creative ideation session (techniques + memlog)
/prd — create, update, or validate a PRD
/architecture-spine — architecture spine
/help — recommends next lifecycle step from your artifacts
Audit / ship:
/secure — security audit
/ship — pre-deploy scorecard
Your workflow:
Tiny changes: just ask, I'll confirm before skipping planning
Small-medium: /write-plan → /execute-plan (wave-based, parallel)
Large/new: /brainstorm → /write-plan → /execute-plan
Lifecycle: /help → /brainstorm-session or /prd → /architecture-spine → /sprint-planning