Project Core Structure
Project architecture contract synced from the project module source documents.
Define the root object chain of project: outcome, workstream, task, activity, and assets.
Explain the main operating loop, five architecture planes, control plane, and execution unit chain.
Define execution mode, runtime, commands, hooks, automation, continuity, and operating posture.
Explain how project change becomes versioned, reviewable, and promotable through git domains.
Map the architecture onto the current board UI, tabs, control plane, and execution projections.
Define what GitHub, Notion, Linear, and Cloudflare own, and where project-level version governance must begin.
Purpose
This document is the board home and architecture index for the project module.
It does not try to explain the whole project system in one uninterrupted essay.
Its job is to define:
- what the
toProjectboard is for - which blocks organize the board
- which topics belong to each block
- which pages define each part of the system
- how the reader should move through the board
This page is the reader's map before the detailed docs begin.
Why This Exists
The project module reached a point where UI refinement alone was no longer enough.
We hit repeated structural drift:
Work itemsdrifting toward timeline/gantt behaviorTimelineandWork itemssharing rendering assumptions that should have stayed separateOverview,Outcome,Workstream,Tasks, andNotescarrying overlapping meaningsProjectitself behaving partly like a task container, partly like a document, and partly like a runtime panel
That made one thing clear:
The missing piece was not another visual pass.
The missing piece was a project-centered architecture contract.
That is why this index exists first.
If we start from isolated sub-documents without restating the argument, the architecture becomes mechanically structured but intellectually thin. The reader can see the folders and headings, but not the reasoning chain that forced the split.
Core Thesis
Project is the git-backed execution container of the agentic platform.
It is not:
- only a runtime panel
- only a task board
- only a document surface
- only a wrapper around an execution engine
It is the root coordination object that organizes:
- outcome
- workstream
- task
- activity
- asset / file
- timeline projection
- runtime state
under an agent-led operating model.
This is the sentence that all other documents are meant to defend, refine, or implement.
Board Definition
toProject is a board, not just a folder of project docs.
As a board, it should do three things at once:
- define the semantic center of project
- explain how project executes
- map that architecture onto the current product surface
That means the board needs internal structure.
The board is not one long article.
It is a system of:
boardblocktopicpage
For toProject, those layers should be read like this:
board:toProjectblock: one major architecture lane inside the boardtopic: one bounded knowledge concern inside a blockpage: the concrete document that explains that topic
Block Structure
The toProject board should be read through five blocks.
Use the cards above as the fast entry path.
Use the sections below as the detailed block map.
1. Core Definition
This block defines what project is.
It owns topics such as:
- root object
- outcome
- workstream
- task
- activity
- assets / files
Primary page:
2. Execution Architecture
This block defines how the whole project system operates as one execution architecture.
It owns topics such as:
- main execution loop
- five architecture planes
- control plane
- tab architecture
- execution unit chain
- failure modes
Primary page:
3. Operating System
This block defines runtime-facing execution behavior.
It owns topics such as:
- execution mode
- runtime
- execution sandbox
- actor model
- capability model
- runtime scheduling
- commands and skills
- hooks and automation
- continuity
- blocking and permission posture
Primary page:
4. Version Spine
This block defines how project change becomes versioned, reviewable, and promotable.
It owns topics such as:
- project-level git role
- spec domain
- execution domain
- knowledge domain
- review boundary
- deploy linkage
Primary page:
5. Surface Mapping
This block defines how the architecture is currently expressed in product UI.
It owns topics such as:
- control plane fields
- tab mapping
- view projections
- missing fields
- transitional surfaces
Primary page:
6. Boundary & Version Strategy
This block defines why toProject exists alongside existing tools, what each external system owns, and how versions are governed across source, execution, and artifacts.
It owns topics such as:
- why
toProjectexists - what GitHub owns
- what Notion owns
- what Linear owns
- what Cloudflare owns
- source version
- execution version
- artifact version
- run ledger
- artifact ledger
- when GitHub is not enough
Primary page:
Topic Map
The board should be read as a topic map, not just as a doc list.
Core Definition Topics
- Project Root
- Outcome
- Workstream
- Task
- Activity
- Assets & Files
Execution Architecture Topics
- Main Loop
- Five Planes
- Control Plane
- Tab Architecture
- Execution Unit Chain
- Execution State Machine
- Observability & Evaluation
- Failure Modes
- Implementation Priorities
Operating System Topics
- Execution Mode
- Runtime
- Execution Sandbox
- Actor Model
- Capability Model
- Runtime Scheduling
- Commands
- Skills
- Hooks
- Automation
- Continuity
- Activity As Operating Evidence
Version Spine Topics
- Why Git Is Core
- Spec Domain
- Execution Domain
- Knowledge Domain
- Review Boundary
- Project-Level Git Surface
Surface Mapping Topics
- Overview Tab
- Outcome Tab
- Workstream Tab
- Tasks Tab
- Activity Tab
- Assets & Files Tab
- List / Board / Timeline Projections
- Current Gaps
Boundary & Version Strategy Topics
- Why
toProjectExists - What GitHub Owns
- What Notion Owns
- What Linear Owns
- What Cloudflare Owns
- Version Management Strategy
- Run Ledger
- Artifact Ledger
- When GitHub Is Not Enough
Document Structure
This architecture is intentionally split into focused documents to avoid one oversized, high-context page.
The split is not for its own sake.
We split because the architecture now has two different reading modes:
- a reader trying to understand the whole project-centered thesis
- a reader trying to answer one bounded question, such as git, runtime, or UI mapping
One document cannot do both well without becoming too long and expensive to read.
1. Index
- PROJECT_CORE_STRUCTURE.md
- Purpose: board home, block map, and reading order
2. Object Model
- PROJECT_OBJECT_MODEL.md
- Focus:
whatthe project-centered object model is andwhyit exists - Reason it is separate: this is the semantic center of the system, and it needs uninterrupted narrative space
3. Git Model
- PROJECT_GIT_MODEL.md
- Focus:
whatgit domains exist, why git is core, and how project binds to them conceptually - Reason it is separate: git is no longer a tooling footnote; it is part of the project architecture itself
4. Runtime Model
- PROJECT_RUNTIME_MODEL.md
- Focus:
whatthe operating layer is, including runtime, execution mode, commands, hooks, automation, and continuity - Reason it is separate: runtime logic is essential, but if mixed into the object model it will overpower the project-centered thesis
5. Execution Architecture
- PROJECT_EXECUTION_ARCHITECTURE.md
- Focus: how intent, decomposition, runtime, git, and evidence combine into one operating loop
- Reason it is separate: this is the synthesis doc that turns the individual definition docs into one readable execution system
6. UI Mapping
- PROJECT_UI_MAPPING.md
- Focus: how the current UI maps to the architecture, what is aligned, and what is still transitional
- Reason it is separate: current implementation state should not dilute the definition docs
Reading Order
Recommended reading order:
- this document
PROJECT_OBJECT_MODEL.mdPROJECT_EXECUTION_ARCHITECTURE.mdPROJECT_RUNTIME_MODEL.mdPROJECT_GIT_MODEL.mdPROJECT_UI_MAPPING.md
That order is intentional:
- object model first, so the root semantics are stable
- execution architecture second, so the reader sees how the parts combine
- runtime and git next, so operating details are interpreted in context
- UI mapping last, so implementation state does not distort definition
Suggested Paths
Use these reading paths depending on the question in front of you.
If the question is semantic
Read:
If the question is operational
Read:
If the question is implementation-facing
Read:
Architecture Shape
The architecture is split into three levels:
1. Project-Centered Definition
This defines the root object and its subordinate object types.
Primary source:
PROJECT_OBJECT_MODEL.md
This level answers:
- what is the root object
- what belongs inside it
- what semantic chain gives the module coherence
2. Operating Definition
This defines how the project is allowed to execute and how versioned execution is governed.
Primary sources:
PROJECT_GIT_MODEL.mdPROJECT_RUNTIME_MODEL.mdPROJECT_EXECUTION_ARCHITECTURE.md
This level answers:
- how the project is versioned
- how it is allowed to run
- how those planes combine into one operating loop
- how engines, commands, hooks, and continuity relate to project semantics
3. UI And Surface Mapping
This defines how the current implementation should express the architecture.
Primary source:
PROJECT_UI_MAPPING.md
This level answers:
- where the current UI already expresses the architecture
- where it is still using transitional placeholders
Reference Position
This architecture is informed by agentic platform engineering references such as the Saul Fernandez model, but it is not a direct copy.
We adopt principles, then extend them locally.
The most important adopted principles are:
- agent intelligence must be treated as infrastructure
- git is the backbone
- intent, execution, and knowledge should be separated
- execution requires explicit operating surfaces
The most important local extensions are:
Projectas the primary product objectOutcome -> Workstream -> Taskas the core structural chainActivityas execution evidence- project-centered UI contracts
- Cloudflare-only runtime assumptions
Current Direction
The current implementation direction should converge toward:
- project as a git-backed execution container
- Cloudflare as the runtime plane
- outcome, workstream, task, activity, and assets as explicit object surfaces
- overview as a summary surface, not a dumping ground
- work items, board, and timeline as separate projections over the same execution model
In other words:
- object semantics must stay stable
- operating semantics must stay explicit
- UI projections must stay subordinate to both
Non-Goals
- putting the entire architecture into one endlessly growing page
- collapsing project into runtime
- collapsing project into task tracking
- collapsing project into documentation only
- collapsing project into a single execution engine
Board Home Use
This page should remain the board home.
It should always answer:
- what
toProjectis for - which blocks exist
- which topics belong to them
- where a reader should go next
Next Docs
The next architecture documents should continue the same split:
- definition docs answer
what + why - implementation docs answer
how
This split is required to keep doc.topoo.ai readable, composable, and low-context.
Execution Architecture Summary
The toProject board should be read as one operating system for delivery.
That means the board is not just where tasks are listed.
It is where five planes are made to work together:
- intent
- decomposition
- operation
- git versioning
- evidence
The canonical chain is:
Project -> Outcome -> Workstream -> Task -> Activity / Assets
The operating loop is:
Intent -> Workstream -> Task -> Runtime -> Git -> Activity -> Updated Project State
That loop is what makes the board more than a planning tool.
It turns the board into the execution architecture of the product.