Skip to content

Latest commit

 

History

12 Commits

Folders and files

NameName
Last commit message
Last commit date
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

Foreman

AI-powered strategic advisor for entrepreneurs.
The frameworks of McKinsey. The philosophy of Marcus Aurelius. The accountability of a great mentor.
Open source. Free forever.

Quickstart · Website · Docs · Vision · Contribute

npm MIT License Stars Issues Visitors Buy Me a Coffee


What is Foreman?

Foreman is an open-source strategic advisor that gives every entrepreneur access to structured strategic thinking, the kind that used to require a six-figure consulting engagement.

You describe your problem in plain language. Foreman compares plausible causes, selects a framework, applies it to your context, and helps produce a deliverable. Confirmed tasks and playbook checkpoints can be saved for your next session.

Without Foreman With Foreman
❌ "My sales are dropping" stays a vague worry ✅ Revenue components compared before attributing a cause
❌ Frameworks learned but never applied ✅ 158 frameworks applied to YOUR context
❌ Board deck built from scratch every time ✅ 48 professional templates, ready to fill
❌ Advice given, execution forgotten ✅ Saved tasks, status history and review prompts
❌ Same strategy for SaaS and marketplace ✅ 9 industry packs with sector-specific benchmarks
❌ Generic AI chat with no memory ✅ Persisted profile, tasks and playbook checkpoints

How It Works

Entrepreneur: "My competitors are crushing us on price"
        │
   ┌────▼────┐
   │  Hooks   │ ── Pattern matches intent → competition threat
   └────┬────┘
   ┌────▼────────────┐
   │  Orchestrator    │ ── Reads memory (Series A, SaaS, 20 employees)
   └────┬────────────┘
   ┌────▼────────────┐
   │  Diagnostic      │ ── 6 triage questions → "differentiation failure"
   └────┬────────────┘
   ┌────▼────────────┐
   │  Playbook        │ ── Competitive Response: Porter's → Art of War → Blue Ocean
   └────┬────────────┘
   ┌────▼────────────┐
   │  Output          │ ── Competitive landscape brief (investor) + action plan (team)
   └────┬────────────┘
   ┌────▼────────────┐
   │  Memory          │ ── Saves confirmed tasks and checkpoints for the next session
   └─────────────────┘

What's Inside

📚 158 Skills 📋 48 Output Templates 🔍 24 Diagnostics
Business frameworks from Porter to Stoic philosophy, across 12 categories Professional documents for investors, boards, teams, clients, and yourself Triage systems that identify root causes through structured questioning
📖 21 Playbooks 🎯 17 Hooks 🤖 6 Agents
Multi-step recipes for pivots, fundraising, scaling, crises, and more Natural language triggers that route problems to the right solution Orchestration layer connecting all components
🏭 9 Industry Packs 🔬 18 Research Guides 🎭 10 Board Personas
Sector-specific benchmarks for SaaS, Marketplace, Fintech, and 6 more Data collection templates from market sizing to churn analysis Adversarial role-play for pitch and board meeting practice

Three Modes

Foreman adapts to how you work:

Mode Command What It Does
Solo /solo Adapts everything for solopreneurs: no team, no board, just you. Remaps audiences, adjusts priorities, acknowledges the isolation.
Stoic /stoic on Adds philosophical depth from Stoic tradition. Dichotomy of control on every analysis. Premeditatio malorum on every risk section. Virtue-based anti-patterns.
Language /language turkish Switches all output to any language. System thinks in English, speaks in yours. Works with every mode.

All three combine: /solo + /stoic on + /language japanese = a solopreneur receiving Stoic-framed strategic guidance in Japanese.


Industry Packs

Same framework, radically different application:

Pack Benchmarks Key Focus
SaaS MRR, ARR, NRR, churn, CAC, LTV Subscription metrics, retention curves
Marketplace GMV, take rate, liquidity, supply/demand Two-sided dynamics, chicken-and-egg
E-Commerce AOV, conversion, cart abandonment, ROAS Physical product, inventory, seasonality
Fintech TPV, default rate, compliance cost, fraud Regulation-through-everything
AI/ML Inference cost, hallucination rate, model drift Model-centric, data moat
HealthTech Clinical validation, provider adoption, HIPAA Patient safety, triple-aim
EdTech Completion rate, institutional adoption, FERPA Multi-stakeholder (buyer ≠ user)
D2C/Consumer CAC by channel, ROAS, brand awareness Brand-as-moat, community
Agency/Consulting Utilization rate, revenue per employee People-as-product

Platform Support

Foreman runs wherever AI agents run:

Platform Integration How
Claude Code Native .claude/ directory: skills, commands, hooks, agents, memory
Claude Code Plugin Marketplace /plugin marketplace add fatihguner/foreman/plugin install foreman@foreman-marketplace
Codex Plugin .codex-plugin/plugin.json + the shared plugins/foreman/skills/ bundle
OpenClaw Native plugin openclaw.plugin.json + compiled dist/index.js registering 10 agent tools

Single source of truth: .claude/. Every other platform derives from it.


Quickstart

Install the stable release

1.1.1 is a maintenance release of the installer, shared state runtime and complete plugin bundles. See the release notes for changes, upgrade steps and limitations. Requires Node 22.19 or later.

npx foreman-sh init

Add --platform codex for Codex, or --platform all for both Claude and Codex. Foreman is MIT-licensed and free to install; your AI host account and model usage are separate.

Use npx foreman-sh@1.1.1 init to pin this version. You can also install the same tarball directly from GitHub:

npx --yes --package=https://github.com/fatihguner/foreman/releases/download/v1.1.1/foreman-sh-1.1.1.tgz foreman-sh init

The installer adds .claude/ and a project CLAUDE.md to your current directory, preserves existing files, prints the next steps and names the one line to add when a CLAUDE.md already exists. npx foreman-sh --version reports the installed version. Repeated initialization only fills missing files; when upgrading from 1.1.0, delete the thirteen command guide files that the old installer placed in .claude/commands/, which the installer lists, and review other generated files before replacing them. Use --workspace /path/to/project to choose a directory or --platform codex to install project skills under .agents/skills/. Open Claude Code and start talking:

"My SaaS churn rate is 5.2% and I'm preparing for Series A. What should I focus on?"

Install as Claude Code plugin

/plugin marketplace add fatihguner/foreman
/plugin install foreman@foreman-marketplace

Clone from GitHub

The CLI runs from a clone without a build step:

git clone https://github.com/fatihguner/foreman.git
cd foreman
npm ci --ignore-scripts
node bin/cli.js init --workspace /path/to/your/project

npm run build compiles the OpenClaw adapter, and npm run build:plugins regenerates the catalog, command wrappers and plugin bundles after content changes. See development and verification.

Codex

npx foreman-sh init --platform codex --workspace /path/to/your/project

This installs the foreman router skill and 158 framework skills under .agents/skills/ with their content under .agents/content/.

OpenClaw

git clone https://github.com/fatihguner/foreman.git
cd foreman
npm ci --ignore-scripts
./scripts/openclaw-setup.sh

Key Commands

These are Claude project commands, generated into .claude/commands/ from the specifications in .claude/command-guides/. Claude plugin commands use /foreman:apply, /foreman:solo, and the same namespace for other commands. Host conflicts use prefixes such as /foreman-run, /foreman-skill, /foreman-resume, /foreman-context and /foreman-help. In Codex, use $foreman followed by the operation. The generated command map is in .claude/catalog.json.

Preferences and implementation tasks persist under .foreman/<founder-id>/state.json through the bundled runtime. The store writes its own .gitignore, so founder records stay out of version control. See runtime behavior and development and verification.

Command What It Does
/skills List all 158 skills, filter by category or stage
/apply [skill] Apply a framework to your context
/diagnose [area] Run a diagnostic triage
/foreman-run [playbook] Start a multi-step playbook
/track View implementation progress
/simulate board Practice your board presentation
/research [topic] Get a data collection guide
/foreman-context View your stored company context
/sector [name] Activate an industry pack
/foreman-help See all commands

Advanced Systems

Implementation Tracking: accountability, not just advice

Every recommendation becomes a trackable item. Six states (not-started → in-progress → blocked → completed → abandoned → deferred), weekly check-ins via /check-in, blocker diagnosis with 10 root causes, and a stuck-protocols library covering all 14 skill categories.

The gap between knowing and doing is where entrepreneurial value dies. Foreman closes it.

Board Simulation: practice before the real meeting

10 adversarial personas (skeptic investor, financial hawk, operational expert, industry veteran, legal counsel, and 5 more) challenge your presentations with hard questions. 5-dimension scoring (clarity, evidence, strategic coherence, risk awareness, response quality). Post-simulation diagnostic identifies your weakest points.

Organizational Politics: navigate the humans, not just the strategy

Stakeholder power mapping, resistance diagnosis, coalition building, communication strategy. Because the best strategy fails if the CFO blocks it and the board is split.

Research Guides: learn to gather your own data

18 structured guides: market sizing (TAM/SAM/SOM), competitor intelligence, customer interviews (34-question bank), pricing validation (Van Westendorp), financial modeling, due diligence prep, and 12 more. Foreman teaches the framework; research guides teach you how to feed it.

5-Layer Memory: a mentor who remembers

Identity (yearly) → Company (monthly) → History (append-only) → Active (weekly) → Session (ephemeral). Your stage, sector, past analyses, ongoing playbooks, and pending decisions persist across conversations. Foreman doesn't start from scratch; it builds on what came before.


Architecture

foreman/
├── .claude/                     # All operational content (single source of truth)
│   ├── skills/                  # 158 skills across 12 categories
│   ├── output-templates/        # 48 templates across 5 audiences
│   ├── diagnostics/             # 24 triage systems
│   ├── playbooks/               # 21 multi-step recipes
│   ├── hooks/                   # 17 natural language triggers
│   ├── agents/                  # 6 orchestration agents
│   ├── memory/                  # 5-layer persistence system
│   ├── command-guides/          # 13 command specifications
│   ├── commands/                # 46 generated command wrappers
│   ├── industry-packs/          # 9 sector-specific overlay packs
│   ├── research/                # 18 data collection guides
│   ├── implementation/          # Tracking + support system
│   ├── simulation/              # Board simulation + 10 personas
│   ├── org-politics/            # Stakeholder navigation
│   ├── solo-mode/               # Solopreneur adaptations
│   ├── stoic-mode/              # Stoic philosophical lens
│   └── language-mode/           # Multilingual output
├── lib/                         # Content, distribution and state runtime modules
├── scripts/                     # 24 shell entry points + Node tooling
├── templates/                   # CLAUDE.md installed into founder projects
├── docs/                        # Architecture, guides, style guide
├── examples/                    # 8 end-to-end walkthroughs
├── openclaw-templates/          # OpenClaw workspace templates
└── .github/                     # CI, issue templates, labels

For the full architecture deep-dive: docs/architecture.md


Skill Categories

Category Skills What It Covers
frameworks 50 Porter's, SWOT, BMC, Blue Ocean, Balanced Scorecard, NPS, and 44 more
leadership 28 Senior transition, delegation, resilience, feedback, mentoring, and more
writing 13 Purpose, process, clarity, tone, emails, memos, performance reviews
ai-leadership 9 AI adoption: learning, purpose, inclusion, vision, augmentation
game-theory 7 Nash equilibrium, prisoner's dilemma, information asymmetry, coalitions
stoic 12 Control, emotions, awareness, duty, resilience, mortality, amor fati
storytelling 8 Finding stories, structure, stakes, emotional impact, performance
negotiation 4 Mindset, competitive tactics, cooperative strategy, closing
people 8 Understanding, meeting, reading, connecting, influencing, conflict
creative 8 Identity, perception, practice, obstacles, flow, collaboration
thinking 6 Clear thinking defaults, strength, weakness, decisions, priorities
decisions 5 Traps, generating solutions, evaluating options, executing, tough calls

Contributing

Foreman is open source because strategic thinking should be open. Every contribution (a new skill, a refined diagnostic, a sector-specific industry pack) helps an entrepreneur somewhere make a better decision.

git clone https://github.com/fatihguner/foreman.git
cd foreman
./scripts/setup.sh

Read the full guide: CONTRIBUTING.md


Roadmap

  • 158 skills across 12 categories
  • 48 output templates across 5 audiences
  • 24 diagnostics + 21 playbooks
  • 9 industry packs with benchmarks
  • Board simulation with 10 personas
  • Implementation tracking + support
  • Solo, Stoic, and Language modes
  • Claude Code + Codex + OpenClaw support
  • Community-contributed skills and packs
  • Anonymous cross-user benchmarks
  • Live data integration (CRM, analytics)
  • Predictive diagnostics
  • Adaptive playbooks

Built By

Fatih Guner, founder of Komunite (Türkiye's largest entrepreneurship community) and Lokomotif AI (corporate AI adoption). Built Foreman after watching thousands of entrepreneurs struggle with the same strategic blind spots.

𝕏 @fatihguner · LinkedIn


License

MIT. Fatih Guner, 2026.


Foreman

Open source under MIT. Built for the 99% of entrepreneurs who will never hire McKinsey, and shouldn't have to.

About

AI-powered strategic advisor for entrepreneurs. 158 skills, 24 diagnostics, 21 playbooks, 9 industry packs. Claude Code + Codex + OpenClaw.

Topics

Resources

Code of conduct

Contributing

Security policy

Stars

51 stars

Watchers

3 watching

Forks

Releases

Packages

Contributors

Languages