Core Principle

Your AI sees everything.
Not just what you said.

In Vibe Coding, AI's context is a black box — you don't know what it assumed. In OpenLogos, every decision lives in a document. Reviewable. Reproducible. Cumulative.

Black box vs. open book

Vibe Coding
You say:
"Build a login feature"
AI thinks (hidden):
Requirements? Guessing...
UI design? Guessing...
API structure? Guessing...
Edge cases? Guessing...
Error handling? Guessing...
Output varies wildly. Every session starts from zero. Decisions are lost.
OpenLogos
You say:
"Build a login feature"
AI reads (explicit):
Requirements doc (Phase 1)
Product design (Phase 2)
Sequence diagram S03
OpenAPI YAML + DB schema
Test cases (12 scenarios)
Output is consistent. Context accumulates. Decisions are preserved.

Everything lives in logos/

A structured directory where every phase's output becomes the next phase's input — all in human-readable formats.

logos/
logos.config.json Project config
logos-project.yaml AI collaboration index
resources/
prd/
1-product-requirements/ Phase 1 · WHY
2-product-design/ Phase 2 · WHAT
1-feature-specs/
2-page-design/
3-technical-plan/ Phase 3 · HOW
1-architecture/
2-scenario-implementation/
api/ OpenAPI YAML specs
database/ SQL DDL / schema
test/ Test case specs (Markdown)
scenario/ API orchestration tests (JSON)
verify/ Acceptance reports
changes/ Delta change proposals
<slug>/ Active proposal
proposal.md Impact analysis + summary
tasks.md Phase-based task checklist
deltas/ Changed artifacts per task
archive/ Completed proposals
skills/ AI Skills (SKILL.md per skill)
prd-writer/
product-designer/
scenario-architect/
code-implementor/
api-designer/ … 13 built-in skills
spec/ Methodology specifications

Three properties that change everything

Reviewable

Every piece of information AI uses is written in Markdown, YAML, or JSON. You can read, audit, and correct any assumption before code is generated.

Reproducible

Same documents → same context → consistent AI output. Switch AI tools, switch team members — the result stays predictable because the context is explicit.

Cumulative

Documents are project knowledge assets. Decisions never get lost. New team members read the docs and understand the full history — no tribal knowledge needed.

AGENTS.md — AI's GPS

When AI opens your project, it reads AGENTS.md first. This file tells it where you are, what to read, and what rules to follow.

AGENTS.md
Phase detection: Scans logos/resources/ to find current phase
Next step: Suggests what to do next based on what's missing
Active skills: Lists which AI Skills to load for each task
Rules: "Never write code without design docs" — enforced automatically
The AI doesn't need you to explain the project every time. It reads the docs, detects the phase, loads the right Skill, and gets to work — with full context.

Explore the foundation

See the engineering theory behind the methodology.