← All runbooks
gooseworks-ai / playbooks-client-packet-engine

Client Packet Engine

This runbook automates a batch GTM pitch-packet pipeline. Given a list of company names and URLs, it runs intelligence gathering and strategy generation (via the client-onboarding playbook), presents ranked strategies for human selection, executes selected strategies in pitch-pac

agent claude-codemodel claude-sonnet-4-6snapshot python312-uveval programmatic8 stepsv1.0.0

Deploy Client Packet Engine to your jetty.io

One-click installs this runbook into a collection on your Jetty account. You can run it from the Spot dashboard, schedule it, or pipe inputs in via the API.

The shape of the run

8 steps · start to finish.

  1. 1
    Step 1

    Environment Setup

    Validate all inputs and confirm prerequisites before processing any companies.

    echo "=== Environment Setup ==="
    
    # Verify required inputs
    if [ -z "$COMPANY_LIST" ]; then
      echo "ERROR: COMPANY_LIST is not set. Provide a YAML list of {name, url} objects."
      exit 1
    fi
    
    # Validate at least 1 company
    COMPANY_COUNT=$(echo "$COMPANY_LIST" | python3 -c "import sys, yaml; data=yaml.safe_load(sys.stdin); print(len(data.get('companies', data)))")
    if [ "$COMPANY_COUNT" -lt 1 ]; then
      echo "ERROR: company_list must contain at least 1 company."
      exit 1
    fi
    
    if [ "$COMPANY_COUNT" -gt 10 ]; then
      echo "WARNING: $COMPANY_COUNT companies in batch — this will be a long run with high cost."
      echo "Estimated cost: ~\$$(echo "$COMPANY_COUNT * 3.5" | bc) total. Proceed? (y/n)"
    fi
    
    # Create output structure
    mkdir -p /app/results/clients
    mkdir -p /app/results/work
    
    echo "Companies to process: $COMPANY_COUNT"
    echo "Budget per company: ${BUDGET_PER_COMPANY:-$5.00}"
    echo "Pitch-packet mode: ${PITCH_PACKET_MODE:-true}"
    echo "Setup complete."
    
  2. 2
    Step 2

    Phase 0 — Intake & Validation

    Parse, validate, and cost-estimate all input companies before any work begins.

  3. 3
    Step 3

    Phase 1 — Research & Strategy Generation (max 3 rounds of retry per company)

    Run the `client-onboarding` playbook (Phases 1–3) for each company, up to **3 companies in parallel**.

  4. 4
    Step 4

    Phase 2 — Strategy Selection (Human Checkpoint)

    Present all generated strategies across all companies for user selection.

  5. 5
    Step 5

    Phase 3 — Strategy Execution (Pitch-Packet Mode)

    Execute selected strategies using the pattern router below. Run up to **5 strategy executions concurrently**. Pause and confirm with user before any single API call >$2.

  6. 6
    Step 6

    Phase 4 — Packaging

    Run the `client-package-local` playbook for each company to produce the final deliverable.

  7. 7
    Step 7

    Iterate on Errors (max 3 rounds)

    For any failed skill execution or packaging step:

  8. 8
    Step 8

    Final Report & Output

    After all companies are packaged, write the batch summary.