Help us improve
Share bugs, ideas, or general feedback.
From maxim
Bootstrap a new project with ARIA scaffolding — project-manifest, watch-profile, brand-foundation, session-memory, operator-profile.
npx claudepluginhub drnabeelkhan/maxim --plugin mxm-pack-l3-4-govtechHow this command is triggered — by the user, by Claude, or both
Slash command
/maxim:mxm-new-projectcommands/The summary Claude sees in its command listing — used to decide when to auto-load this command
# /mxm-new-project
## Usage
- Claude Code: `/mxm-new-project`
- Claude CLI: `claude "/mxm-new-project"`
- Claude Desktop: type `/mxm-new-project` in chat
Interactive project scaffold. Creates a new project folder with all Maxim per-project
files, placeholder templates, and `.claude-sessions-memory/` session storage.
Works from Claude Code Desktop chat or CLI — no PowerShell required.
**Triggers:** "new project", "create project", "scaffold project", "maxim new project",
"set up new project", "bootstrap project"
**Requires:** Maxim globally installed (`%USERPROFILE%\.claude\...Share bugs, ideas, or general feedback.
/mxm-new-projectclaude "/mxm-new-project"/mxm-new-project in chatInteractive project scaffold. Creates a new project folder with all Maxim per-project
files, placeholder templates, and .claude-sessions-memory/ session storage.
Works from Claude Code Desktop chat or CLI — no PowerShell required.
Triggers: "new project", "create project", "scaffold project", "maxim new project",
"set up new project", "bootstrap project"
Requires: Maxim globally installed (%USERPROFILE%\.claude\CLAUDE.md exists)
Ask these questions one at a time in sequence. Wait for each answer before asking the next:
Project folder name
What should the project folder be named?
(e.g. fixit-leads, drivingtutors-ca, gulflaw-ai)
Parent location
Where should the folder be created?
(e.g. E:\Projects or ~/projects — press Enter for current directory)
Project display name
What is the full project name?
(e.g. FixIt Leads Platform)
Vertical
What vertical is this project?
SaaS | AI | FinTech | HealthTech | LegalTech | EdTech | Other
Compliance markets
Which markets does this project target? (select all that apply)
CA (PIPEDA) | US (TCPA) | EU (GDPR) | UAE/Gulf (UAE-PDPL)
Payments (PCI-DSS) | Health (HIPAA) | Enterprise (SOC2) | None yet
Stage
What stage is this project?
idea | early-stage | active | scaling | mature
Output a summary and ask for confirmation:
┌─────────────────────────────────────────────────────┐
│ Maxim New Project — Confirm │
├─────────────────────────────────────────────────────┤
│ Folder : {parent}/{folder-name}/ │
│ Name : {display-name} │
│ Vertical : {vertical} │
│ Compliance : {frameworks} │
│ Stage : {stage} │
│ │
│ Will create 19 files + 7 folders. │
│ Proceed? (yes / no) │
└─────────────────────────────────────────────────────┘
Create the following directories inside {parent}/{folder-name}/:
{folder-name}/
├── config/
├── documents/
│ ├── ADRs/ ← Architecture Decision Records (gitignored, private ledger)
│ │ ├── INDEX.md ← auto-created: empty ADR index
│ │ └── TEMPLATE.md ← auto-created: ADR template
│ ├── architecture/ ← PRD, FRD, SRD, ARCHITECTURE.md (gitignored)
│ │ └── .secrets/ ← build intakes, API keys, credentials (gitignored)
│ └── business/ ← investor narrative, financial models (gitignored)
├── prototypes/ ← v0 demos, POCs, throwaway builds
├── .mxm-skills/
├── .mxm-operator-profile/
└── .claude-sessions-memory/
documents/ADRs/ is the canonical location Maxim uses for every new project's ADR ledger. The folder is created on bootstrap, seeded with a placeholder INDEX.md and TEMPLATE.md, and is gitignored by default — ADRs are an operator-local private ledger until explicitly published.
Seed documents/architecture/ with templates from templates/project-scaffold/documents/architecture/ (ARCHITECTURE.md, PRD.md, FRD.md, SRD.md, README.md).
Create each file with the content specified. Substitute user inputs where shown.
config/project-manifest.jsonCopy content from Maxim source config/project-manifest.TEMPLATE.json and replace:
"your-project-id" → "{folder-name}""Your Project Name" → "{display-name}""your-github-username" → read from config/project-manifest.json at Maxim root if available, else leave placeholder"vertical" → "{vertical}""stage" → "{stage}"compliance.frameworks[] → array of selected framework strings"YYYY-MM-DD" (last_updated) → today's dateconfig/agent-registry.jsonCopy from Maxim source config/agent-registry.json unchanged.
CLAUDE.project.mdCopy content from Maxim source documents/templates/CLAUDE.project.TEMPLATE.md and replace:
<!-- REQUIRED: Your project name --> → {display-name}your-project-id → {folder-name}Your Project Name → {display-name}SaaS | AI | FinTech... → {vertical}PIPEDA | GDPR... → selected compliance frameworks (comma-separated)<!-- REQUIRED: YYYY-MM-DD --> → today's date.claude.local/settings.local.json{
"permissions": {
"allow": [
"Bash(git *)",
"Bash(node -e \":*)",
"Bash(python3 -c \":*)",
"Read({full-project-path}/**)"
]
}
}
Replace {full-project-path} with the resolved absolute path using forward slashes.
.mxm-skills/agents-skill-gaps.log# Maxim Skill Gaps Log — {display-name}
# Format: [ISO-TIMESTAMP] [SKILL-GAP|Maxim-UNENHANCED] domain: message
# Review after every session. Entries here drive skill priority.
.mxm-skills/agents-background.log# Maxim Background Agent Log — {display-name}
# Format: [ISO-TIMESTAMP] [BACKGROUND-AGENT|BACKGROUND-FAIL] message
.mxm-skills/agents-handoff.md# Maxim Agent Handoff — {display-name}
# Written by orchestrators during multi-agent task chains.
# This file is runtime-only — do not commit.
.claude-sessions-memory/handoff.md# Session Handoff — {display-name}
# Last updated: {today}
# Use this file to resume context across Claude Code sessions.
## Last Session Summary
<!-- Claude Code writes here automatically at session end -->
## Open Tasks
<!-- Incomplete items from last session -->
## Key Decisions Made
<!-- Architecture or product decisions logged here -->
## Next Session Start Point
<!-- What to load and run first next session -->
.claude-sessions-memory/decision-log.md# Decision Log — {display-name}
# Format: [DATE] [DECISION] rationale
# Append entries. Never delete.
.claude-sessions-memory/skill-gaps.log# Session Skill Gaps — {display-name}
# Format: [ISO-TIMESTAMP] [SKILL-GAP|Maxim-UNENHANCED] domain: message
.gitignore# ── Maxim Session Memory (not for version control) ──────────────────────────
.claude-sessions-memory/
# ── Maxim Operator Profile (personal, never commit) ─────────────────────────
.mxm-operator-profile/
# ── Maxim Runtime State ────────────────────────────────────────────────────
.mxm-skills/agents-handoff.md
.mxm-skills/agents-skill-gaps.log
.mxm-skills/agents-background.log
# ── Maxim Local Config ────────────────────────────────────────────────────
.claude.local/
config/project-manifest.local.json
# ── Architecture & secrets (project-specific, never commit) ──────────────
documents/architecture/
# ── MemPalace (local data, never commit) ─────────────────────────────────
config/mempalace.yaml
.mempalace/
# ── Generated ────────────────────────────────────────────────────────────
GEMINI.md
documents/reference/AGENTS.md
.github/copilot-instructions.md
# ── Common ───────────────────────────────────────────────────────────────
.env
.env.local
node_modules/
__pycache__/
*.pyc
.DS_Store
Thumbs.db
.mxm-operator-profile/OPERATOR.md# Operator Profile
# Updated automatically at session end by Maxim behavioral-designer agent.
# Manual edits welcome — Maxim will preserve and build on them.
## Identity
<!-- Name, role, organization -->
## Working Style
<!-- How you prefer to work -->
## Output Preferences
<!-- Preferred response format -->
## Technical Context
<!-- Primary languages, frameworks, tools -->
.mxm-operator-profile/preferences.md# Operator Preferences
# Append-only. Maxim adds entries as it learns your preferences.
.mxm-operator-profile/rejected-patterns.md# Rejected Patterns
# Patterns you have corrected or rejected. Maxim will avoid these.
# NEVER delete entries — only add.
.mxm-operator-profile/communication-style.md# Communication Style
# How you phrase requests, preferred response format, verbosity level.
✅ Maxim Project Created
────────────────────────────────────────────────────────
Folder : {full-project-path}/
Files : 15 files created
Memory : .claude-sessions-memory/ (session storage)
Registry : config/agent-registry.json ✅
Manifest : config/project-manifest.json ✅ ← fill in REQUIRED fields
Project : CLAUDE.project.md ✅ ← fill in brand voice + rules
Gitignore: .gitignore ✅ (.claude-sessions-memory/ excluded)
────────────────────────────────────────────────────────
Next steps:
1. Open {full-project-path}/ in Claude Code
2. Run: /mxm-status
3. Fill in CLAUDE.project.md brand voice section
4. Fill in config/project-manifest.json REQUIRED fields
────────────────────────────────────────────────────────
After creation, verify all 15 files exist. Report any missing as ❌ MISSING with
a one-line fix command. If all present, output 🟢 HIGH — project scaffold complete.