Skip to main content

Documentation Index

Fetch the complete documentation index at: https://ade-app.dev/docs/llms.txt

Use this file to discover all available pages before exploring further.

​
What is the CTO Agent?

The CTO (Chief Technical Officer) is ADE’s persistent, project-scoped AI agent — always running in the background watching commits, monitoring lane status, tracking PR progress, and accumulating project understanding over time. Its system prompt includes a baked-in Memory Protocol (search before acting, save findings immediately), a Daily Context routine (check recent decisions and worker activity at each activation), and a Decision Framework (act autonomously when safe, escalate when risky). These protocols survive compaction.

Always-On Awareness

Runs on a configurable heartbeat, checking for new commits, PR status changes, Linear issue updates, and stuck lanes — even when you are not looking.

Org Chart Model

The CTO sits at the top of an agent hierarchy. It spawns and coordinates specialized Worker agents (Employees) to handle specific tasks.

Bidirectional Linear Sync

Reads from and writes to Linear. Creates issues from code changes, updates issue status when PRs merge, and adds analysis comments automatically.

PR Convergence

Drives PRs to merge readiness autonomously — starts convergence rounds, monitors CI and review comments, and manages the full resolution loop via five dedicated operator tools.

Deep Memory

Maintains multi-scope persistent memory across project, identity, daily-log, run, and runtime-thread scopes. The CTO remembers what happened yesterday.

​
CTO vs. Regular Agents

Understanding how the CTO differs from the agents you spawn manually (in chat sessions or missions) helps you know when to use each.
CTO AgentRegular / Worker Agent
LifecycleAlways on; persistent across sessionsEphemeral; lives for one task or session
ScopeEntire projectSpecific lane, task, or mission
MemoryAccumulates over weeks/monthsWithin the current session or run
TriggersHeartbeat schedule + events + manual chatExplicit user action or automation trigger
AuthorityCan spawn lanes, create missions, delegate to workers, drive PR convergenceOperates within its assigned lane/context
Linear accessFull bidirectional sync (owner)Read-only via ADE CLI tools, if granted
Primary UICTO tabLane terminal, Chat pane, Missions tab

​
The Org Chart Model

The CTO implements an org chart model for AI work delegation. 
CTO (always-on, project-scoped)
├── Worker: "auth-specialist"   (assigned to lane/auth-refactor)
├── Worker: "test-writer"       (ephemeral, created for one task)
└── Worker: "pr-reviewer"       (persistent, reviews all new PRs)
The CTO sits at the top. Below it are Worker agents (also called Employees) — specialized sub-agents the CTO creates and assigns to specific tasks or lanes. Workers have:
  • A name and specialization prompt that defines their role and constraints
  • An assigned lane (optional; some workers operate across lanes)
  • A tool palette — a subset of ADE CLI tools they are permitted to use
  • A budget limit — a token/cost cap enforced by the CTO
The CTO coordinates workers: it assigns tasks, monitors their output, routes their results back to Linear or PRs, and intervenes if a worker is stuck or over budget.

​
Explore CTO Features

Setting Up the CTO

Walk through onboarding, heartbeat configuration, and verifying the CTO is active.

CTO Memory

Understand multi-scope persistent memory, memory tiers, and the memory review surface.

Linear Integration

Connect Linear for bidirectional issue sync, status updates, and intelligent triage.

Workers & Budget

Hire workers, manage budgets, configure practical workflows, and troubleshoot issues.