A working operating system for a one-person business — strategy, time telemetry, and AI agent operations, published as reusable docs, templates, and automation. 혼자 일하지만, 시스템은 팀처럼 굴리는 방법.
Solo Biz Playbook is the actual playbook BuildnWrite uses to run a solo business, anonymized for public use. It connects three things most solopreneur resources treat separately: a one-page strategy (Lean Canvas), measured time telemetry (automatic worklogs), and model-neutral AI agent operations (Claude Code, Codex, or any CLI agent). Copy the templates, run the automation, and adapt the decision principles to your own business.
KO 이 저장소는 BuildnWrite가 실제 1인 사업 운영에 쓰는 문서와 자동화를 공개 가능한 형태로 익명화해 정리한 플레이북입니다. 전략(린캔버스), 시간 계측(워크로그), AI 에이전트 운영 규칙을 하나의 운영 시스템으로 연결합니다.
When you work alone, you own strategy, execution, records, automation, and retrospectives all at once. Most solopreneur material stops at "how to validate an idea." It rarely connects the idea to the operating layer: how you measure your time, how you keep AI tooling costs accountable, and how you keep your workflows portable when models and CLIs change.
This repository fills that gap. Set direction with a Lean Canvas, measure time with automatic worklogs, run AI agents on a model-neutral document structure, and review your AI subscription usage every month.
KO 대부분의 자료는 아이디어 검증에서 멈춥니다. 이 플레이북은 그 다음, 즉 실제 운영에 필요한 계측·자동화·AI 도구 사용 규칙까지 연결하는 공백을 메웁니다.
- Solopreneurs, freelancers, and creators who handle sales, production, and operations alone
- One-person teams inside a company owning data, automation, education, content, or ops improvement
- Heavy AI-tool users who want to track where their output, time, cost, and learning actually accumulate
- Anyone who wants to compound personal know-how into documents, templates, and automation
- A one-page strategy canvas to describe your business (with a real, anonymized example)
- A worklog structure that records measured time per project, automatically
- Agent operating rules that survive switching between Claude, Codex, or any other AI agent
- A monthly review report format for your AI subscription usage (tokens, cost, categories, models)
- Copy-ready templates and publicly safe, anonymized real-world examples
Follow the four steps in order. Each step has a blank template you copy and fill in.
Close the 5-axis naming constitution and tone signature before generating candidate names. Generating candidates too early makes the evaluation criteria drift and postpones the decision. Module: docs/brand-naming/ · Template: template/brand-naming-brief.md
KO 후보 단어를 너무 일찍 만들면 평가 기준이 매번 흔들리고 결정이 미뤄집니다. 5축 헌법과 톤 시그니처를 먼저 닫는 것이 핵심입니다.
Write a Lean Canvas that states your customer, problem, solution, and revenue structure. Render the finished canvas as a single image and check for strategy drift every quarter. Module: docs/strategy-design/ · Template: template/lean-canvas.md
Treat these three modules as one system. Drop any one of them and you will re-make the same judgment call next quarter.
docs/principles/— recurring decision rules: Pricing Floor, Platform Concentration, Wealth Ladderdocs/office-hours/— a self-mentoring procedure that stress-tests a specific decision with forced questions and 2–3 alternativesautomation/claude-worklog/— automatic per-project time measurement (the measured evidence behind the principles)docs/agent-systems/— run Claude, Codex, and other AI agents on one model-neutral structure
Aggregate your Claude subscription usage on every billing date: cost, tokens, work categories, and model mix. The report shows which friction to cut and which high-leverage patterns to keep next month. Module: docs/claude-monthly-review/
KO 청구서만 봐서는 구독 요금제의 실사용 가치를 알기 어렵습니다. 월간 집계를 하면 다음 달에 줄일 마찰과 유지할 고효율 패턴이 보입니다.
| Module | Landing page | What it does |
|---|---|---|
| Strategy Design | docs/strategy-design/ |
Define your customer, problem, and revenue structure with a real example plus a blank template |
| Brand Naming | docs/brand-naming/ |
A 5-step process that closes the 5-axis constitution and tone signature before candidate names |
| Decision Principles | docs/principles/ |
Recurring judgment rules: Pricing Floor, Platform Concentration, Wealth Ladder, Security Incident Response |
| Office Hours | docs/office-hours/ |
Self-mentoring with forced questions: verify assumptions, derive 2–3 alternatives and next actions |
| Operations Telemetry | docs/operations-telemetry/ |
Log every AI agent session to calendar + worklog for measured per-project time |
| Claude Monthly Review | docs/claude-monthly-review/ |
Auto-aggregate monthly AI subscription usage into cost/token/category/model insights |
| Agent Systems | docs/agent-systems/ |
Shared docs, scripts, and thin adapters that keep workflows model-neutral |
| Layer | Role |
|---|---|
docs/ |
Principles, explanations, and architecture diagrams |
examples/ |
Anonymized public versions of real usage |
template/ |
Blank forms readers copy and fill in |
automation/ |
Runnable automation examples and scripts |
Absolute values in cost structure, the Profit First split, and the asset roadmap are withheld; the shares and multiples around them are published. Ratios, strategy, principles, and slogans are intact. Full text:
examples/my-canvas.md· image source:examples/my-canvas.html
The Lean Canvas is one module of the starter kit: study the real example, project your own business onto the blank template, then render the result as a one-page image to check strategy drift.
docs/strategy-design/— module landing pageexamples/my-canvas.md— a real, filled-in canvastemplate/lean-canvas.md— the blank template you copydocs/principles/— recurring decision rules extracted from the canvas
Principles only hold up when backed by measured data. To say "I don't take work below my hourly floor" as a number rather than a slogan, you need telemetry that accumulates per-project hours automatically.
docs/operations-telemetry/— module landing pageautomation/claude-worklog/— a Stop-hook system that writes every Claude Code / Codex session to Google Calendar and an Obsidian worklog- Telemetry architecture diagram (HTML) — how session logs become metrics, calendar entries, and worklogs
- At quarterly review, filter cumulative hours per project as the measured evidence for Principle 01 (Pricing Floor)
KO 원칙은 실측 데이터가 있어야 방어됩니다. 세션이 끝날 때마다 캘린더와 업무일지에 자동 기록이 쌓이고, 분기 회고 때 프로젝트별 누적 시간을 시급 방어선의 근거로 사용합니다.
A flat-rate AI subscription hides its real value in the invoice. Aggregating tokens, cost, sessions, hours, and project distribution each month shows which friction to cut and which usage patterns pay for themselves.
%%{init: {'theme':'base','themeVariables':{'fontFamily':'Pretendard, Apple SD Gothic Neo, sans-serif','fontSize':'13px','primaryColor':'#ffffff','primaryTextColor':'#31302e','primaryBorderColor':'#615d59','lineColor':'#615d59','secondaryColor':'#f6f5f4','tertiaryColor':'#f6f5f4','edgeLabelBackground':'#f6f5f4','clusterBkg':'#f6f5f4','clusterBorder':'#dddddd'}}}%%
flowchart LR
A[AI tool usage] --> B[Monthly aggregation]
B --> C[Cost, tokens, hours]
B --> D[Qualitative insights]
C --> E[Next month's actions]
D --> E
classDef focal fill:#E9EFFD,stroke:#2563EB,stroke-width:1.5px,color:#1D4ED8;
class E focal;
docs/claude-monthly-review/— module landing pageexamples/monthly-claude-review/— an anonymized real monthly report
AI tools keep changing. If your workflow knowledge lives only in one model's prompt file, you rebuild the same automation every time you switch models or CLIs.
docs/agent-systems/— model-neutral agent operation docsdocs/agent-systems/agent-guide.md— shared SSOT docs with thin per-agent adaptersdocs/agent-systems/utilities-registry.md— a registry template for reusable scripts, templates, and skills- Agent system architecture diagram (HTML) — how shared docs/scripts relate to per-model adapters
KO 업무 시스템을 특정 모델의 프롬프트 파일에만 두면 모델을 바꿀 때마다 같은 자동화를 다시 만들게 됩니다. 공용 문서와 스크립트를 SSOT로 두고, 모델별 어댑터는 얇게 유지하는 구조입니다.
A text canvas is information-dense but hard to take in at a glance. Rendering the finished canvas as one image helps you:
- Skim it weekly and catch strategy drift early
- Attach it directly to meetings and pitches
- Get higher response rates when asking peers or mentors for feedback
In Claude Code (or any capable agent), ask: "Visualize my Lean Canvas as a 9-block grid layout, using my brand colors, as a single self-contained HTML file with inline SVG." You get one HTML file you can keep editing as text.
| Tool | Strength | Weakness |
|---|---|---|
| Notion database board | Easy block editing and linking | Layout breaks when exporting to image |
| Excalidraw | Hand-drawn tone, free-form | No Notion embed; separate file to share |
| Figma / FigJam | Best for collaborative feedback | Overkill for a solo operator |
| Official Lean Canvas (leanstack.com) | The original format | Hard to customize branding |
KO 개인 경험상 Notion에 텍스트 SSOT를 두고 분기마다 한 장짜리 PNG를 내보내는 조합이 유지비가 가장 낮았습니다.
The GitHub front page is the full map. When sharing on Threads or a blog, link the module README that matches the post topic so readers land in the right place.
| Topic you're sharing | Direct link |
|---|---|
| Strategy design | docs/strategy-design/ |
| Brand naming process | docs/brand-naming/ |
| Solo business decision principles | docs/principles/ |
| Self-mentoring for business decisions (Office Hours) | docs/office-hours/ |
| Operations telemetry | docs/operations-telemetry/ |
| Monthly AI subscription review | docs/claude-monthly-review/ |
| Agent operating system | docs/agent-systems/ |
| Real examples only | examples/ |
| Copy-ready templates only | template/ |
| Runnable automation only | automation/ |
Per-directory README conventions live in docs/publishing-structure.md.
Jay Lim — solo operator behind BuildnWrite, working at the intersection of AI × data × education. 30+ corporate lectures, 400+ mentees.
- Threads: @buildnwrite
- LinkedIn: jayjunglim
- Book: n8n Does It All 『n8n이 다해줌』 (Hanbit Media, in Korean)
The kit expands in stages based on reader response. Full plan: ROADMAP.md
Next up:
- v1.1 — practice forms for the principles (screening worksheet, contract clauses, email templates, quarterly retro agenda)
- v1.2 — real decision log (ADR) + change history
- v1.3 — Profit First and KPI tracking Google Sheets links
- License: CC BY-SA 4.0 — fork and remix welcome; attribute and share alike
- Contributing: CONTRIBUTING.md — PRs with your own anonymized Lean Canvas are welcome. The long-term goal is a public collection of solo-business canvases.
KO 자기 린캔버스를 익명화해 PR로 보내주시면 환영합니다. 1인 사업자 캔버스 컬렉션으로 키우는 것이 장기 목표입니다.