Workspace Intelligence knowledge portal

Learn Workspace Intelligence.

Understand how projects, dependencies, models, graphs, evidence, and agent context fit together through practical guides, contracts, and reproducible examples.

Start learningBrowse concepts

A simple mental model

From workspace inputs to shared understanding

  1. 1Input LayerProjects · Repositories · Dependencies · Changes
  2. 2Workspace IntelligenceModel · Graph · Impact · Verify
  3. 3Agent GroundingAGENTS.md · Skills · Context
  4. 4Evidence ContractsReports · Artifacts · Gates
  5. 5Shared ConsumersDevelopers · CI · IDEs · AI agents

Learn by outcome

Start with the problem you are trying to solve

Graphs, skills, Doctor, repair, and project kits are not separate buzzwords. Each is a view or workflow over the same Workspace Intelligence foundation.

Give my agents better context

Agent context and skills

Ground Codex, Claude, Copilot, Cursor, Gemini, and other coding agents in the same current workspace evidence.

Learn this path

Show me how everything connects

Code and workspace graph

Explore projects, APIs, dependencies, ownership, proofs, and likely change impact without loading the entire workspace into every prompt.

Learn this path

Find what blocks my project

Workspace Doctor

Detect dependency, security, configuration, test, runtime, and release-readiness problems with project and workspace scope.

Learn this path

Fix it and prove the result

Verified AI repair

Turn a blocker or engineering goal into a scoped repair path that stays open through install, test, build, and final verification.

Learn this path

Start an AI product

AI-ready kits and modules

Start with a production-shaped project and add agent runtime, RAG, vector, LLM gateway, guardrail, authentication, billing, and observability modules.

Learn this path

Give me the complete system

Full Workspace Intelligence

Use the model, graph, Doctor, evidence, verified goals, agent grounding, kits, and modules as one contract-backed operating layer.

Learn this path

Practical entry path

Start with evidence, not a blank prompt

Move from the concept to a machine-readable workspace artifact and verify its contract before any consumer relies on it. Each step is documented against the current CLI command surface.

  1. 01

    Learn the system model

    Start with the boundary between repository files and an evidence-backed model of the software system.

    Understand the concept
  2. 02

    Run the canonical chain

    Execute the versioned dependency graph as one evidence-coherent operation and inspect its authoritative run report.

    Open the intelligence runner
  3. 03

    Verify the boundary

    Validate contracts before downstream tools, CI, IDEs, or agents consume workspace intelligence.

    Open contract verification

Concrete example

Follow a fact from command to governed artifact

A repository scan is only an input. Workspai turns observed facts into named artifacts with explicit schema versions, then exposes those artifacts to humans and agents through documented command boundaries.

CommandGoverned artifactSchema version
workspai workspace intelligence run --for-agent generic --strict --json.workspai/reports/workspace-intelligence-run-last-run.jsonworkspace-intelligence-run.v1
workspai workspace model --json --write.workspai/reports/workspace-model.jsonworkspace-model.v1
workspai workspace graph search "authentication endpoint" --limit 12 --json.workspai/reports/workspace-knowledge-graph.jsonworkspace-knowledge-graph.v1
workspai workspace contract verify --strict --json.workspai/reports/workspace-contract-verify-last-run.jsonworkspace-contract-verify.v1
workspai workspace context --for-agent --json --write.workspai/reports/workspace-context-agent.jsonworkspace-context.v1

Category boundary

Workspace Intelligence is more than retrieval

Indexing, retrieval, and memory remain valuable inputs. Workspace Intelligence adds a shared system model, evidence semantics, versioned outputs, and verification boundaries that multiple consumers can independently inspect.

CapabilityPrimary questionDurable outputTrust boundary
Repository indexingWhere is matching code or text?Search indexSource presence
RAGWhat retrieved context may answer this prompt?Prompt contextRetrieval relevance
Agent memoryWhat did an agent previously observe or decide?Session or long-term memoryRecorded agent experience
Workspace IntelligenceWhat is this software system, and what evidence supports it?Versioned model, graph, evidence, and decisionsExplicit contracts and provenance

Knowledge map

Explore the complete system

What Is Workspace Intelligence?

Define the evidence-backed understanding layer for software systems.

Concepts

Separate Workspace Intelligence from RAG, memory, repository indexing, skills, and chat.

Software System Understanding

Learn the structure, semantics, relationships, intent, evolution, trust, and reasoning layers.

Mental Models

Reason about shared system models, evidence-backed context, workflows, and agent authority.

Architecture

Map the input layer, intelligence loop, evidence contracts, and consumers that share one workspace truth.

Workspace Model

Study the canonical model of projects, runtimes, commands, policies, contracts, and evidence.

Workspace Graph

Understand the graph of the software system: structure, runtime, ownership, change, evidence, and agents.

Evidence

Learn how Workspai treats verified, observed, inferred, stale, and unknown facts.

Contracts

Inspect every synced contract, field, artifact, producer command, and consumer boundary.

CLI Reference

Inspect every Workspai command and its contract-backed artifact relationships.

Guides

Adopt existing repos, create supported projects, generate agent context, and run verification gates.

RFCs

Track progressive graph layers, Atlas, package separation, and future consumer surfaces without confusing them with shipped baselines.

Glossary

Keep the vocabulary consistent across Workspai, RapidKit, docs, CLI, VS Code, and future cloud surfaces.

Essays

Read category-shaping positions on AI engineering, repositories, and shared software-system truth.

Research

Use durable evaluation frames and architectural requirements for Workspace Intelligence systems.