GOHIGHLEVEL ANSWERS & PROMPTS: PLAYBOOK T21 · MULTI-AGENT HANDOFFS & MCP
AUTHOR: DANIEL HALLMAN
KEYWORD: AGENT HANDOFF
PLAYBOOK T21 · Multi-Agent Handoffs & MCP KEYWORD: AGENT HANDOFF

How Do You Coordinate Codex, Claude Code, and a Planner Agent Without Copy-Pasting?

Answer: Stop copying prompts between chat windows by giving your planner, builder (Codex), and reviewer (Claude Code CLI) a shared task JSON queue in git worktrees plus a scoped GoHighLevel Private Integration MCP token. The planner writes a task spec file, a runner script invokes Codex in an isolated branch, Claude Code reviews the git diff against acceptance criteria, and approved CRM updates execute directly through GoHighLevel MCP.

Tags: #mcp #ai #cli #workflows #webhooks #private-integrations
STEP-BY-STEP INSTRUCTIONS + ANNOTATED GHL SCREENSHOTS

Steps in GoHighLevel

  1. STEP 1
    In Settings > Private Integrations, create a scoped token for your sub-account so your planner and builder CLI agents can query live pipelines, contacts, and workflows via MCP instead of asking you to paste screenshots.
    REAL GHL SCREEN · SETTINGS > PRIVATE INTEGRATIONS (API V2) ANNOTATED GHL SCREENSHOT
    Annotated GoHighLevel Private Integrations API v2.0 screenshot showing Settings > Private Integrations menu, Create New Integration button, and scoped token audit row
    1Open Settings > Private Integrations 2Click Create New Integration 3Review Scoped Token & Last Used Date
  2. STEP 2
    In AI Agents > Knowledge Base (and your repo's tasks/ directory), store shared JSON task specs containing the goal, target files, and acceptance checks so the planner writes the spec file and Codex builds in an isolated git worktree.
    REAL GHL SCREEN · AI AGENTS > KNOWLEDGE BASE ANNOTATED GHL SCREENSHOT
    Annotated GoHighLevel AI Agents Knowledge Base screenshot showing Knowledge Base tab, + Create knowledge base button, and KB gaps table
    1Select Knowledge Base Tab 2Click + Create Knowledge Base 3Review Industry FAQ & Check KB Gaps
  3. STEP 3
    In Automation > Workflows > Execution Logs, trigger a reviewer command (claude -p) on the branch diff before merge and log completed task status back to the opportunity or webhook log.
    REAL GHL SCREEN · WORKFLOW > EXECUTION LOGS ANNOTATED GHL SCREENSHOT
    Annotated GoHighLevel Workflow Execution Logs screenshot showing Execution logs tab, All actions and All statuses filters, and populated contact run rows
    1Open Execution Logs Tab 2Filter by Action Type or Status 3Search Contact to Trace Runs
MODE: AGENT PROMPT
PASTE INTO CLAUDE CODE, CURSOR, CHATGPT, OR GOHIGHLEVEL ASK AI AGENT PROMPT 
Connect planner, builder, and reviewer CLI agents to sub-account <YOUR_LOCATION_ID>:
1. In Settings > Private Integrations, create a scoped token with Contacts, Opportunities, Calendars, and Workflows scopes.
2. Have the planner write task specifications to tasks/<task_id>.json with explicit acceptance checks.
3. Run the builder agent in an isolated git worktree, run the reviewer CLI against git diff, and log completion webhooks to Automation > Workflows.
Cause
  • Running a planner model in a browser chat, copying its prompt into Codex, pasting Codex's output into Claude Code for review, and copying review notes back turns the operator into a manual clipboard.
  • As soon as you get busy on client calls, the handoff stalls and context drifts between windows.
GoHighLevel Configuration
  • Replace chat-window copy-pasting with a file-backed task spec (tasks/<id>.json), isolated git worktrees per task, an automated git diff review gate from a second model family
  • and a scoped GoHighLevel MCP connection for live CRM state.
MENU PATH REFERENCE

GoHighLevel Menu Paths

Step Menu Path in GHL Action
Step 1 Settings > Private Integrations Create a scoped token for your sub-account so your planner and builder CLI agents can query live pipelines, contacts, and workflows via MCP instead of asking you to paste screenshots.
Step 2 AI Agents > Knowledge Base Store shared JSON task specs containing the goal, target files, and acceptance checks so the planner writes the spec file and Codex builds in an isolated git worktree.
Step 3 Automation > Workflows > Execution Logs Trigger a reviewer command (claude -p) on the branch diff before merge and log completed task status back to the opportunity or webhook log.
3-step diagram showing Shared Spec task JSON written by planner, Branch Build by Codex in an isolated worktree, and Diff Gate checked by reviewer before merge.
VISUAL PLAYBOOK · PLAYBOOK T21 (MULTI-AGENT HANDOFFS & MCP) Back to Annotated GHL Screenshots ↑
DEFAULT SETUP VS. AUTOMATED WORKFLOW

Operational Comparison

Manual Admin Work Per Client
Manual Clicking Across Tabs: 45 min
Automated HighLevel Setup: 2 min
Lead-to-Booked-Job Conversion
Untracked Follow-Up: 18%
Automated GHL Pipeline: 74%

Related Questions

BUILD FASTER IN GOHIGHLEVEL · FREE TRIAL BONUSES + MCP SERVER

Build this directly inside your GoHighLevel sub-account

Connect Claude Code, Cursor, or OpenAI Codex to our ghl-tools MCP Server & Rust CLI to build workflows, forms, and funnels from your editor, or start your GoHighLevel plan through our partner link to unlock $612+ in ready-to-import kits (including GHL MCP & Multi-Agent Harness Pack) and weekly live clinics.

Explore ghl-tools MCP Server → Join Operator Community & Claim Bonus Vault Start 14-Day GHL Trial + Free $47 Kit Get GHL MCP & Multi-Agent Harness Pack