The Smallest "AI Company"
That Can Ship

jaan.to handles the repetitive work—formatting PRDs, writing boilerplate, generating checklists.
You focus on what matters: defining the right problem, making trade-offs, understanding customers.

60-80% First Draft Time Saved
40% More Time for Decisions
60% Test Planning Automated
MIT Open Source

The Problem

The same request produces different outcomes depending on how you ask. Most people blame the model. The harder truth: humans are the inconsistent part.

📝

Without Systems

Request a feature → one architectural style. Request a bug fix → different style. Rules in your head. Standards that only appear when you get frustrated. A definition of "done" that changes with your energy level.

With jaan.to

Rules written down, not in your head. Standards that exist before frustration. A definition of "done" that doesn't change with energy levels.

The Four Pillars

The problem isn't AI. The problem is system design.

📝

Commands, not prompts

A command is not a "nice prompt." It's a procedure with predictable output.

👥

Roles, not cleverness

Real teams don't have one person do everything. Different roles create different thinking.

📚

Memory, not vibes

If rules live in your head, you don't have memory. You have vibes.

🛡

Gates, not hope

People skip gates because nobody is watching. Production is always watching.

This Is Not Replacing Teams

jaan.to standardizes execution and reduces waste. The repetitive parts—formatting PRDs, writing test matrices, generating boilerplate—get automated.

The human parts stay human:

🎯

Clarity

Defining the right problem to solve

Judgment

Making trade-off decisions

Empathy

Understanding real customer needs

Quality

Knowing when "done" means done

We measure impact and iterate like a product. If a skill isn't helping, we fix it or remove it.

Usage Guide

Everything you need to use jaan.to. No fluff.

Skills Reference

Skills are slash commands that generate outputs. Each skill follows the two-phase workflow with human approval gates. 37 total skills across 7 roles.

PM (Product Manager)

CommandDescriptionStatus
/jaan-to-pm-prd-writeGenerate Product Requirements DocumentsAvailable
/pm-decision-brief1-page decision record with options and risksComing Soon
/pm-north-starNorth star metric with drivers and boundariesComing Soon
/pm-scope-sliceMVP vs Later slicing with milestonesComing Soon
/pm-release-reviewPost-release review with KPI deltasComing Soon

UX (User Experience)

CommandDescriptionStatus
/ux-flow-specFlow spec with states and edge casesComing Soon
/ux-microcopy-writeMicrocopy pack with labels and errorsComing Soon
/ux-research-planResearch plan with method and timelineComing Soon
/ux-heuristic-reviewHeuristic review with severity and fixesComing Soon
/ux-competitive-reviewCompetitive teardown with opportunitiesComing Soon

DEV (Development)

CommandDescriptionStatus
/dev-tech-planTechnical approach with architecture and risksComing Soon
/dev-test-planDev-owned test plan with fixtures and mocksComing Soon
/dev-api-contractOpenAPI contract with payloads and errorsComing Soon
/dev-pr-reviewPR review pack with risky files and hintsComing Soon
/dev-ship-checkPre-ship checklist with Go/No-GoComing Soon

QA (Quality Assurance)

CommandDescriptionStatus
/qa-test-matrixRisk-based matrix across states and devicesComing Soon
/qa-bug-triageDedupe, severity, and repro hintsComing Soon
/qa-automation-planWhat to automate now vs laterComing Soon
/qa-regression-runbookStep-by-step regression runbookComing Soon
/qa-release-signoffGo/No-Go summary with evidenceComing Soon

DATA (Analytics)

CommandDescriptionStatus
/jaan-to-data-gtm-datalayerGenerate GTM tracking codeAvailable
/data-event-specGA4-ready event/param specificationComing Soon
/data-metric-specMetric definition with formulaComing Soon
/data-funnel-reviewFunnel baseline with hypothesesComing Soon
/data-experiment-designExperiment plan with success metricsComing Soon

SEO (Growth)

CommandDescriptionStatus
/growth-content-outlineWriting-ready outline with H1-H3Coming Soon
/growth-keyword-briefKeyword + intent map with targetsComing Soon
/growth-seo-auditOn-page checklist with content gapsComing Soon
/growth-seo-checkTechnical audit with crawl signalsComing Soon
/growth-weekly-reportWeekly wins/losses with actionsComing Soon

AIOS (System)

CommandDescriptionStatus
/to-jaan-skill-createCreate new skills with wizardAvailable
/to-jaan-skill-updateUpdate existing skillsAvailable
/to-jaan-docs-createCreate documentation with templatesAvailable
/to-jaan-docs-updateAudit and maintain documentationAvailable
/to-jaan-learn-addAdd lessons to improve skillsAvailable

Two-Phase Workflow

Every skill follows this workflow pattern to ensure quality and maintain human oversight:

Phase 1: Analysis (Read-Only)

  1. Load Context: Read skill definition, LEARN.md, context
  2. Interview: Ask clarifying questions
  3. Plan: Structure the output mentally
# Skill asks 3 questions: 1. What problem does this solve for users? 2. How will you measure success? 3. What is explicitly NOT in scope?

Phase 2: Generation (After Approval)

  1. Generate: Create artifact from template
  2. Preview: Show complete output to user
  3. Write: Save to .jaan-to/outputs/ on approval
# HARD STOP: Human approval required Ready to proceed? [y/n] # After approval, writes to: .jaan-to/outputs/pm/spec/{slug}/prd.md

Configuration

jaan.to uses context to inject team and tech context into skills:

Stack Files

# Tech context - languages, frameworks, tools jaan-to/context/tech.md # Team context - structure, ceremonies, norms jaan-to/context/team.md # Integration context - Jira, GitLab, Slack APIs jaan-to/context/integrations.md # Main configuration jaan-to/config.md
Markdown-based No JSON schemas Easy to edit

Learning System

Every skill improves from feedback through the LEARN.md system:

How Learning Works

Run skill → Get output → Give feedback → Next run is better

4 Categories:
  • Better Questions
  • Edge Cases
  • Workflow
  • Common Mistakes

Three Learning Layers

# Skill learning (execution) .claude/skills/{skill}/LEARN.md # Template learning (output format) jaan-to/templates/LEARN.md # Stack learning (context) jaan-to/context/LEARN.md

Safety & Guardrails

jaan.to has non-negotiable safety rules to protect your work:

Always Enforced

  • ✔ Safe paths only - writes to .jaan-to/
  • ✔ Preview first - see before saving
  • ✔ Approval required - explicit user consent
  • ✔ No credentials - scanned for secrets

Denied Paths

  • src/, lib/ - source code
  • .env - environment files
  • *.config.* - config files
  • package.json - dependencies

Skills

Commands remove ambiguity. Ambiguity is where AI starts guessing. Guessing is where production debt is born.

Product Manager Skills

1 available, 4 coming soon
📝

Decision Brief

Coming Soon

1-page decision record with options, recommendation, and risks. Perfect for stakeholder alignment.

/pm-decision-brief
🎯

North Star

Coming Soon

Define north star metric with drivers, boundaries, and review cadence. Connects to GA4 for baselines.

/pm-north-star

Scope Slice

Coming Soon

MVP vs Later slicing with milestones and dependencies. Integrates with Jira and GitLab.

/pm-scope-slice
📈

Release Review

Coming Soon

Post-release review with KPI deltas, learnings, and follow-ups. Connects to GA4, Clarity, Sentry.

/pm-release-review

User Experience Skills

0 available, 5 coming soon
🛠

Flow Spec

Coming Soon

Flow spec with happy path, empty/loading/error states, and edge cases. Reads from Figma.

/ux-flow-spec
💬

Microcopy Write

Coming Soon

Microcopy pack with labels, helper text, errors, toasts, and empty states. Figma integration.

/ux-microcopy-write
🔍

Research Plan

Coming Soon

Research plan with questions, method, participants, script, and timeline.

/ux-research-plan
🔎

Heuristic Review

Coming Soon

Heuristic review with issues, severity, fixes, quick wins, and principles.

/ux-heuristic-review
📊

Competitive Review

Coming Soon

Competitive teardown with flows, patterns, strengths/weaknesses, and opportunities.

/ux-competitive-review

Development Skills

0 available, 5 coming soon
💻

Tech Plan

Coming Soon

Technical approach with architecture, tradeoffs, risks, and rollout/rollback strategy.

/dev-tech-plan
🔧

Test Plan

Coming Soon

Dev-owned test plan with unit/integration/e2e scope, fixtures, and mocks.

/dev-test-plan
🔗

API Contract

Coming Soon

OpenAPI contract with payloads, errors, versioning, and examples.

/dev-api-contract
🔎

PR Review

Coming Soon

PR review pack with summary, risky files, security/perf hints, and missing tests.

/dev-pr-review
🚀

Ship Check

Coming Soon

Pre-ship checklist with flags, migrations, monitoring, rollback, and Go/No-Go.

/dev-ship-check

Quality Assurance Skills

0 available, 5 coming soon
📋

Test Matrix

Coming Soon

Risk-based matrix with P0/P1 flows across states, devices, and environments.

/qa-test-matrix
🐛

Bug Triage

Coming Soon

Dedupe, severity, repro hints, next action. Clusters by root cause.

/qa-bug-triage
🤖

Automation Plan

Coming Soon

What to automate now vs later, flakiness risk, and testability changes.

/qa-automation-plan
📖

Regression Runbook

Coming Soon

Step-by-step regression runbook from smoke to critical to deep checks.

/qa-regression-runbook

Release Signoff

Coming Soon

Go/No-Go summary with evidence, open risks, mitigations, and rollback readiness.

/qa-release-signoff

Analytics & Data Skills

1 available, 5 coming soon
📈

Event Spec

Coming Soon

GA4-ready event/param spec with naming, triggers, and required properties.

/data-event-spec
📉

Metric Spec

Coming Soon

Metric definition with formula, caveats, segmentation rules, and owner.

/data-metric-spec
📐

Funnel Review

Coming Soon

Funnel baseline with top drop-offs, segments, and 3-5 hypotheses.

/data-funnel-review
🔬

Experiment Design

Coming Soon

Experiment plan with hypothesis, success metric, boundaries, and ramp/kill criteria.

/data-experiment-design
🚨

Anomaly Triage

Coming Soon

Triage pack with scope, causes, next checks, who to pull in, and RCA template.

/data-anomaly-triage

SEO & Growth Skills

0 available, 5 coming soon
📝

Content Outline

Coming Soon

Writing-ready outline with H1-H3, FAQs, entities, and internal links.

/growth-content-outline
🔑

Keyword Brief

Coming Soon

Keyword + intent map with primary/secondary targets and SERP notes.

/growth-keyword-brief
🔍

SEO Audit

Coming Soon

On-page checklist with title/meta, headings, content gaps, and links.

/growth-seo-audit
🔧

SEO Check

Coming Soon

Technical audit with indexability, crawl signals, and critical errors.

/growth-seo-check
📈

Weekly Report

Coming Soon

Weekly wins/losses, top pages/queries, actions with owners, and experiments.

/growth-weekly-report

System & Internal Skills

6 available, 0 coming soon

37 Total Skills across 7 roles. 8 available now, 29 coming in Phase 3.

View Full Roadmap

Setup

Two paths: quick start or full configuration. Both work.

Quick Setup

Clone the repository and start using skills immediately:

# Clone the repository git clone https://github.com/jaan-to/claude-code.git cd ai-os-claude # Open in Claude Code claude . # Run your first skill /jaan-to-pm-prd-write "your feature idea"
Recommended 5 minutes No config required

Full Setup

Configure context and customize for your team:

# Configure your tech stack vim jaan-to/context/tech.md # Add languages, frameworks, infrastructure # Configure your team vim jaan-to/context/team.md # Add team structure, ceremonies, norms # Configure integrations (optional) vim jaan-to/context/integrations.md
Team Context Custom Stacks MCP Ready

System Requirements

Required

  • ✔ Claude Code CLI
  • ✔ Git 2.20+
  • ✔ Claude API access

Recommended

  • ✔ VS Code or Cursor
  • ✔ Configured context
  • ✔ LEARN.md for feedback

Optional

  • ◯ MCP connectors
  • ◯ Custom hooks
  • ◯ Custom skills

Human-Centered by Design

This is not replacing teams. This is making teams faster.
40-60% of product work is repetitive. We automate that. You keep the decisions.

🐙

GitHub Repository

Explore the source code, report issues, and contribute to the project. All skills and documentation are open source.

View on GitHub
📖

Documentation

Comprehensive documentation covering all skills, workflows, and configuration. Start with the getting-started guide.

Read Docs
📍

Roadmap

See what's coming next for jaan.to. Track development progress and upcoming features in the roadmap.

View Roadmap

One Truth

"Humans become more senior when AI handles execution.
Not replaced. Elevated."

🧑

Humans decide

AI executes

Judgment stays

Boilerplate goes

📈

Teams grow

Not shrink

The Full Story

Read "The Smallest AI Company That Can Ship" on Medium for the story behind jaan.to.
Compare notes on AI systems that actually ship.

PK

Parhum Khoshbakht

Creator