Guides¶

Pick by what you are doing¶

You want to…	Guide
Run only part of coding — or none of the Docker parts	Composing What Runs
Add a new coding agent	Agent Integration
Work behind a corporate proxy or VPN	Network Configuration
Decode the tmux status bar	Status Line
Investigate a health problem	Health Dashboard
Add, switch or debug an LLM provider	LLM Providers
Write or debug a constraint	Constraint Testing
Explore the knowledge graph	VKB Visualization
Understand how knowledge is captured	Knowledge Workflows
Work out why CI went red	Continuous Integration
Write or use a skill	Skills System

If you are new¶

Read Getting Started first, then Core Systems. The guides here go deeper than either and assume you already have something running.

What these guides are for¶

The Core Systems pages explain what each system is. These guides are about doing something with them — adding an agent, chasing a failure, changing a provider — and so carry complete configurations and worked examples rather than overviews.

Choosing one¶

Guide	Reach for it when
Composing What Runs	Deciding which of the nine features to run — presets, dependencies, and what a pared-down install costs you
Agent Integration	Adding a coding assistant to the system — the config contract, hooks and API
Network Configuration	Behind a corporate proxy or VPN, or an agent's traffic is being routed unexpectedly
Status Line	Decoding the tmux bar — every indicator, and how to configure it
Health Dashboard	A service is unhealthy, or you want to understand the supervision architecture
LLM Providers	Adding, switching or troubleshooting a cloud or local provider
Constraint Testing	Writing a constraint, or one is firing when it should not
VKB Visualization	Exploring the knowledge graph, or driving the viewer programmatically
Knowledge Workflows	Understanding how knowledge is captured, processed and stored
Continuous Integration	A CI run went red, or you want to know what a green one actually proves
Skills System	Writing a skill, or working out why one did not trigger

Two that pay off early¶

Status Line. It is on screen the entire time you work, and it reports health, cost and logging state continuously. Learning to read it turns most "is something broken?" questions into a glance.

Health Dashboard. When something is broken, this is where the answer is. It also explains the layered supervision model, which is what makes an unresponsive-but-alive process detectable at all.

Related¶

• Getting Started — installation and first session
• Core Systems — what each system does
• Architecture — why it is built this way
• Troubleshooting — symptom-first index

Comprehensive reference guides with detailed explanations, complete configurations, and visual walkthroughs.

When to Use These Guides¶

Guide	Use When...
Agent Integration	Adding a new AI coding assistant to the system
Status Line	You need to understand status bar indicators
Health Dashboard	Investigating health issues or understanding the monitoring architecture
Constraint Testing	Adding new constraints, debugging detection, or testing enforcement
VKB Visualization	Exploring knowledge graphs or integrating VKB programmatically
Knowledge Workflows	Understanding how knowledge is captured, processed, and stored
Continuous Integration	A CI run went red, or you need to know what a green one actually proves
Network Configuration	Working behind a corporate proxy or VPN, or routing an agent's API traffic
LLM Providers	Adding, switching, or troubleshooting a cloud or local LLM provider

Quick Links¶

• Getting Started - Initial setup and configuration
• Core Systems - Overview of LSL, UKB/VKB, Constraints
• Architecture - System architecture and data flow
• Troubleshooting - Common issues and solutions