/prime

description: Prime agent with full Archon codebase understanding

Prime: Load Archon Project Context

Objective

Build comprehensive understanding of the Archon monorepo before beginning any work. This orients you on structure, active state, and key architectural conventions.

Process

1. Analyze Project Structure

List the monorepo packages: !`ls packages/`

Show workspace scripts: !`cat package.json | head -35`

2. Read Core Documentation

Read `AGENTS.md` in full for project rules and development guidelines. Follow its links to `.archon/direction.md`, `.archon/engineering.md`, and the docs for focused project context.

3. Identify Key Entry Points

Read these files to understand the primary message and execution flow:

• `packages/core/src/orchestrator/orchestrator-agent.ts` — single entry point for all platforms
• `packages/core/src/handlers/command-handler.ts` — slash command routing (no AI)
• `packages/server/src/index.ts` — server startup, adapter wiring, port allocation
• `packages/workflows/src/executor.ts` — run setup; every workflow executes as a DAG in `dag-executor.ts`

4. Understand Package Dependency Layers

Inspect the relevant `packages/*/package.json` manifests for workspace dependencies before reporting dependency order. Read each package's entry points to understand its responsibility.

5. Understand Current State

Check recent commits: !`git log -10 --oneline`

Check current branch and working tree status: !`git status`

List any active worktrees: !`git worktree list`

Output Report

Provide a concise summary (under 300 words) covering:

Project Overview

• Archon: Self-hostable, governed agentic automation engine
• Bun + TypeScript monorepo; use `packages/` and the root `package.json` for the current workspace structure
• SQLite (default, zero-setup) or PostgreSQL (optional via DATABASE_URL)

Architecture

• Package dependency order and each package's responsibility
• Key interfaces: `IPlatformAdapter`, `IAgentProvider`, `IDatabase`, `IWorkflowStore`
• Message flow: platform adapter → orchestrator-agent → command handler OR AI provider
• Workflow execution: `discoverWorkflows` → router → `executeWorkflow` → `executeDagWorkflow`

Current State

• Active branch, recent changes, any uncommitted work
• Any observations relevant to next task

**Keep it scannable — bullets over prose.**