Give it a goal. The whole company works until it's done.

/company "Build the user auth system with OAuth2"

A Claude Code skill that reads your team structure from COMPANY.md, runs every employee in loops, and doesn't stop until built-in reviewers verify the goal is met.

Install

npx company-skill install

Or from git:

curl -sL https://raw.githubusercontent.com/jagmarques/company-skill/main/install.sh | bash

Edit COMPANY.md with your team. Or skip it, the skill creates a minimal company automatically.

How It Works

GOAL: "Build the auth system"

  THINK     CEO picks relevant employees, leads assign tasks
  EXECUTE   Employees do the work, use installed skills
  VERIFY    Reviewer checks criteria.json, Advocate attacks results

  Not done? Loop back with feedback.
  Done? Update playbook, write STATUS.md.

The loop runs until ALL criteria in criteria.json pass. A Stop Hook blocks Claude from exiting early. To cancel: touch .company/CANCEL.

Goal Enforcement

The skill creates criteria.json with checkable success criteria:

{"goal":"Build auth","criteria":[
  {"id":1,"description":"OAuth2 login works with Google","passes":false,"evidence":null},
  {"id":2,"description":"All tests pass","passes":false,"evidence":null}
]}

The reviewer updates passes to true with evidence as work completes. The stop hook reads this file and blocks exit until everything passes.

Self-Improvement

One file: .company/playbook.md. Accumulates across sessions.

After each session, the CEO writes:

• WORKED: what succeeded (with evidence)
• FAILED: what failed, USE INSTEAD: what works, WHY: the difference
• INEFFICIENT: what was slow, FASTER: better approach
• TOP: best employees for priority activation next time
• HIRE/FIRE: roles added or deactivated

Leads read the playbook before every THINK phase. Employees check failed approaches before proposing new ones. The company that starts session 5 is smarter than session 1.

The CEO also updates COMPANY.md: tags [inactive] on zero-contribution roles, [priority] on top performers, adds hired roles, evolves employee descriptions based on what they're good at.

Built-In Roles

Every company gets these automatically:

Role	Phase	What they do
CEO	THINK	Picks relevant employees for the goal, resolves conflicts
CTO	THINK	Technical decisions, architecture
Internal Reviewer	VERIFY	Checks criteria.json, rejects findings without sources
User Advocate	VERIFY	Represents the end user
Devil's Advocate	VERIFY	Attacks results, finds holes
Elegance Enforcer	VERIFY	Prevents over-engineering

Deduplicated if you define them in COMPANY.md.

Source Citations

Every finding needs a source:

• Existing claims: file path, URL, or command output
• Novel ideas: "NOVEL - needs validation" (reviewer adds a validation criterion)

No source = rejected by reviewer.

Commands

/company "Build X"      Run until X is done
/company                Run using COMPANY.md priorities
/company:run "Build X"  Same as above
/company:status         Show last status
/company:resume         Continue from last session

Visual Indicators

════════════════════════════════════════════════
CYCLE 1 - THINK > EXECUTE > VERIFY
════════════════════════════════════════════════

CYCLE 1 VERDICT: NOT DONE
Missing validation of compression ratios

Employees show with colors: leads (cyan), workers (green), reviewers (yellow), digest (gray).

Agents

Agent	Phase	Color
company-lead	THINK	Cyan
company-worker	EXECUTE	Green
company-reviewer	VERIFY	Yellow
company-critic	VERIFY	Yellow
company-digest	COMPRESS	Gray

Model Assignment

Phase	Model	Who
THINK	Opus	CEO, CTO, leads
EXECUTE	Sonnet	Workers
VERIFY	Opus	Reviewers
COMPRESS	Haiku	Digest writer

Override per employee: - ML Scientist, experiments [opus]

Installed Skills

Auto-installed on first run:

Pack	What employees get
gstack	/review, /ship, /qa, /investigate, /browse, /office-hours
GSD	/gsd:plan-phase, /gsd:execute-phase, /gsd:verify-work, /gsd:debug
trailofbits	Security audit, vulnerability detection

Install manually for more:

/plugin marketplace add obra/superpowers-marketplace
/plugin marketplace add wshobson/agents
/plugin marketplace add alirezarezvani/claude-skills

When installed, employees MUST use them.

What Gets Created

.company/
  criteria.json        Machine-checkable goal state
  playbook.md          Accumulated lessons (self-improvement)
  active-roster.md     Employees activated for this goal
  active-tasks.md      Deduplicated task list
  STATUS.md            Final report
  cycles/              Per-cycle briefings and reviews
  messages/            Typed findings per department
  {dept}/              Per-employee findings (persist across sessions)

Examples

File	Team
startup.md	10-person startup
research-lab.md	Academic group
dev-team.md	Dev sprint
nexusquant.md	Full research company

License

MIT