Clawdtalk Client

ClawdTalk

Voice calling, SMS messaging, and AI Missions for Clawdbot. Call your bot by phone, send texts, or run autonomous multi-step outreach campaigns — powered by ClawdTalk.

Trust: By using this skill, voice transcripts, SMS messages, and mission data are sent to clawdtalk.com (operated by Telnyx). Only install if you trust this service with your conversation data.

External Endpoints

Security & Privacy

• Voice transcripts and SMS content are transmitted to clawdtalk.com.
• Mission state and events are stored server-side for tracking and insights.
• setup.sh reads gateway config to extract connection details; with confirmation it adds a voice agent and sessions_send to gateway.tools.allow.
• API key is stored in skill-config.json — use env var CLAWDTALK_API_KEY or a ${CLAWDTALK_API_KEY} reference to avoid plaintext storage.

⚠️ CRITICAL: SLUG CONSISTENCY

init auto-generates a slug from the mission name (lowercased, spaces → hyphens). Every command that takes a slug (setup-agent, save-memory, complete) MUST use the EXACT same slug.

Mismatched slugs = agent not linked = scheduled events invisible on the frontend.

# After init, ALWAYS confirm the slug:
python scripts/telnyx_api.py list-state
# Output: find-window-washing-contractors: Find window washing contractors [running]
#         ^^^^^^^^^^^^^^^^^^^^^^^^^^^^ COPY-PASTE THIS. NEVER ABBREVIATE.

⚠️ CRITICAL: YOU OWN THE MISSION LIFECYCLE

The server does NOT automatically update plan steps or mission status. That is YOUR job as the bot. If you don't update steps and complete the mission, the UI will show "Running" forever with all steps "Pending".

Your Responsibilities

For every mission, YOU must:

1. Update plan steps — mark each step in_progress → completed (or failed)
2. Log events — after every significant action
3. Poll for completion — check if scheduled calls/SMS finished
4. Complete the mission — mark the run as succeeded or failed
5. Clean up — remove any polling cron jobs when the mission ends

The UI reflects exactly what you tell it. No updates from you = no updates on screen.

🚨 MANDATORY: SAVE TO MEMORY AFTER EVERY ACTION

Every significant action MUST be persisted using save-memory or append-memory IMMEDIATELY after the action succeeds. The frontend reads from server memory. If you don't save it, it doesn't show up. log-event alone is NOT enough.

Rule: If you did something, save it to memory. No exceptions. No "I'll do it later." Do it NOW.

Example (scheduling an SMS):

# 1. Schedule it
python scripts/telnyx_api.py schedule-sms $AID "$TO" "$FROM" "$DATETIME" "$MESSAGE" $MID $RID $STEP_ID

# 2. IMMEDIATELY save to memory
python scripts/telnyx_api.py append-memory "$SLUG" "scheduled_events" \
  '{"event_id": "<id>", "type": "sms", "to": "<to>", "message": "<msg>", "scheduled_at": "<dt>", "step_id": "<step>"}'

# 3. Then log the event
python scripts/telnyx_api.py log-event $MID $RID custom "Scheduled SMS event_id=<id>" $STEP_ID

Skipping step 2 is the #1 cause of "nothing shows on the frontend" bugs.

🚨 MANDATORY: CHECK MISSION STATUS AFTER EVERY STEP

After completing or failing ANY step, you MUST check whether the mission should be completed or failed. Never leave a mission in "running" state when it's actually done or dead.

Rule: After every step change, ask yourself: is this mission finished?

Decision tree (run after EVERY step):

Step finished →
  ├── Succeeded?
  │     ├── All steps done? → COMPLETE MISSION (update-run succeeded)
  │     ├── More steps remain? → Continue to next step
  │     └── Only verify left? → Set up polling cron
  └── Failed?
        ├── Recoverable (retry/reschedule)? → Retry
        └── Unrecoverable? → FAIL MISSION NOW:
              1. update-step <step_id> failed
              2. log-event error "Failed: <reason>" <step_id>
              3. save-memory "$SLUG" "error_<step_id>" '{"error": "...", "recoverable": false}'
              4. update-run $MID $RID failed
              5. save-memory "$SLUG" "result" '{"status": "failed", "reason": "...", "failed_step": "..."}'
              6. Clean up any polling cron jobs

A mission stuck in "running" when it's actually done or dead is a bug. The user sees it and thinks work is still happening.

Architecture Overview

How It Fits Together

┌──────────────┐     ┌──────────────────┐     ┌──────────────┐
│  You (Bot)   │────▶│  ClawdTalk Server │────▶│  Telnyx API  │
│              │     │  (dev/prod)       │     │  (cloud)     │
│ telnyx_api.py│     │  Local DB + proxy │     │  Executes    │
│              │     │  to Telnyx        │     │  calls/SMS   │
└──────────────┘     └──────────────────┘     └──────────────┘

• telnyx_api.py — Python CLI script. Every command you run goes through this. It talks to the ClawdTalk server, never directly to Telnyx.
• ClawdTalk Server — Node.js backend. Stores missions, assistants, and events in a local Postgres DB. Proxies requests to the Telnyx API.
• Telnyx API — Cloud service. Actually makes the calls and sends the SMS at the scheduled time.

Key Files

Two Sets of IDs

Every entity exists in two places with different IDs:

• Local DB ID — what the ClawdTalk server returns (e.g. 3df24dde-...)
• Telnyx ID — what the Telnyx API uses internally

The script always works with local IDs. You don't need to worry about Telnyx IDs.

Quick Start

1. Sign up at clawdtalk.com
2. Add your phone in Settings
3. Get API key from Dashboard
4. Run setup: ./setup.sh

   setup.sh reads your gateway config to extract connection details, adds a voice agent to agents.list, and (with confirmation) adds sessions_send to gateway.tools.allow. Gateway config is at ~/.openclaw/openclaw.json or ~/.clawdbot/clawdbot.json.

5. Start connection: ./scripts/connect.sh start

Voice Calls

The WebSocket client routes calls to your gateway's main agent session, giving full access to memory, tools, and context.

./scripts/connect.sh start     # Start connection
./scripts/connect.sh stop      # Stop
./scripts/connect.sh status    # Check status

Outbound Calls

Have the bot call you or others:

./scripts/call.sh                              # Call your phone
./scripts/call.sh "Hey, what's up?"            # Call with greeting
./scripts/call.sh --to +15551234567            # Call external number*
./scripts/call.sh --to +15551234567 "Hello!"   # External with greeting
./scripts/call.sh status <call_id>             # Check call status
./scripts/call.sh end <call_id>                # End call

*External calls require a paid account with a dedicated number. The AI will operate in privacy mode when calling external numbers (won't reveal your private info).

SMS

Send and receive text messages:

./scripts/sms.sh send +15551234567 "Hello!"
./scripts/sms.sh list
./scripts/sms.sh conversations

AI Missions (Full Tracking via Python)

For complex, multi-step missions with full tracking, state persistence, retries, and conversation insights, use the Python-based missions API.

Required: Python 3.7+, CLAWDTALK_API_KEY environment variable. Optionally set CLAWDTALK_API_URL to override the default endpoint (defaults to https://clawdtalk.com/v1).

python scripts/telnyx_api.py check-key    # Verify setup

CRITICAL: SAVE STATE FREQUENTLY

You MUST save your progress after EVERY significant action. If the session crashes or restarts, unsaved work is LOST.

Two-Layer Persistence: Memory + Events

Always save to BOTH:

1. Local Memory (.missions_state.json) - Fast, survives restarts
2. Events API (cloud) - Permanent audit trail, survives local file loss

When to Save (After EVERY action!)

Memory Commands (Local Backup)

# Save a single value
python scripts/telnyx_api.py save-memory "<slug>" "key" '{"data": "value"}'

# Append to a list (great for collecting multiple items)
python scripts/telnyx_api.py append-memory "<slug>" "contractors" '{"name": "ABC Co", "phone": "+1234567890"}'

# Retrieve memory
python scripts/telnyx_api.py get-memory "<slug>"           # Get all memory
python scripts/telnyx_api.py get-memory "<slug>" "key"     # Get specific key

Event Commands (Cloud Backup)

# Log an event (step_id is REQUIRED - links event to a plan step)
python scripts/telnyx_api.py log-event <mission_id> <run_id> <type> "<summary>" <step_id> '[payload_json]'

# Event types: tool_call, custom, message, error, step_started, step_completed
# step_id: Use the step_id from your plan (e.g., "research", "setup", "calls")
#          Use "-" if event doesn't belong to a specific step

When to Use Missions

This skill has two modes: full missions (tracked, multi-step) and simple calls (one-off, no mission overhead). Pick the right one.

Use a Full Mission When:

• The task involves multiple calls or SMS (batch outreach, surveys, sweeps)
• You need a complete audit trail with events, plans, and state tracking
• The task is multi-step and takes significant effort across phases
• Retries and failure tracking matter
• You need to compare results across multiple calls

Examples:

• "Find me window washing contractors in Chicago, call them and negotiate rates"
• "Contact all leads in this list and schedule demos"
• "Call 10 weather stations and find the hottest one"

Do NOT Use a Mission When:

• The task is a single outbound call — just create an assistant (or reuse one) and schedule the call directly
• It's a one-off SMS — schedule it and done
• The task doesn't need tracking, plans, or state recovery
• You'd be creating a mission with one step and one call — that's overengineering

For simple calls, just:

# Reuse or create an assistant
python scripts/telnyx_api.py list-assistants --name=<relevant>
# Schedule the call
python scripts/telnyx_api.py schedule-call <assistant_id> <to> <from> <datetime> <mission_id> <run_id>
# Poll for completion
python scripts/telnyx_api.py get-event <assistant_id> <event_id>
# Get insights
python scripts/telnyx_api.py get-insights <conversation_id>

No mission, no run, no plan. Keep it simple.

State Persistence

The script automatically manages state in .missions_state.json. This survives restarts and supports multiple concurrent missions.

python scripts/telnyx_api.py list-state                              # List all active missions
python scripts/telnyx_api.py get-state "find-window-washing-contractors"  # Get state for specific mission
python scripts/telnyx_api.py remove-state "find-window-washing-contractors" # Remove mission from state

Core Workflow

Phase 1: Initialize Tracking

Step 1.1: Create a Mission

python scripts/telnyx_api.py create-mission "Brief descriptive name" "Full description of the task"

Save the returned mission_id - you'll need it for all subsequent calls.

Step 1.2: Start a Run

python scripts/telnyx_api.py create-run <mission_id> '{"original_request": "The exact user request", "context": "Any relevant context"}'

Save the returned run_id.

Step 1.3: Create a Plan

Before executing, outline your plan:

python scripts/telnyx_api.py create-plan <mission_id> <run_id> '[
  {"step_id": "step_1", "description": "Research contractors online", "sequence": 1},
  {"step_id": "step_2", "description": "Create voice agent for calls", "sequence": 2},
  {"step_id": "step_3", "description": "Schedule calls to each contractor", "sequence": 3},
  {"step_id": "step_4", "description": "Monitor call completions", "sequence": 4},
  {"step_id": "step_5", "description": "Analyze results and select best options", "sequence": 5}
]'

Step 1.4: Set Run to Running

python scripts/telnyx_api.py update-run <mission_id> <run_id> running

High-Level Alternative: Initialize Everything at Once

Use the init command to create mission, run, plan, and set status in one step:

python scripts/telnyx_api.py init "Find window washing contractors" "Find contractors in Chicago, call them, negotiate rates" "User wants window washing quotes" '[
  {"step_id": "research", "description": "Find contractors online", "sequence": 1},
  {"step_id": "setup", "description": "Create voice agent", "sequence": 2},
  {"step_id": "calls", "description": "Schedule and make calls", "sequence": 3},
  {"step_id": "analyze", "description": "Analyze results", "sequence": 4}
]'

This also automatically resumes if a mission with the same name already exists.

⚠️ Immediately after init, run list-state and copy the exact slug. Use it for ALL subsequent commands.

Phase 2: Voice/SMS Agent Setup

When your task requires making calls or sending SMS, create an AI assistant first.

Step 2.1: Create a Voice/SMS Assistant

For phone calls:

python scripts/telnyx_api.py create-assistant "Contractor Outreach Agent" "You are calling on behalf of [COMPANY]. Your goal is to [SPECIFIC GOAL]. Be professional and concise. Collect: [WHAT TO COLLECT]. If they cannot talk now, ask for a good callback time." "Hi, this is an AI assistant calling on behalf of [COMPANY]. Is this [BUSINESS NAME]? I am calling to inquire about your services. Do you have a moment?" '["telephony", "messaging"]'

For SMS:

python scripts/telnyx_api.py create-assistant "SMS Outreach Agent" "You send SMS messages to collect information. Keep messages brief and professional." "Hi! I am reaching out on behalf of [COMPANY] regarding [PURPOSE]. Could you please reply with [REQUESTED INFO]?" '["telephony", "messaging"]'

Save the returned assistant_id.

Step 2.2: Find and Assign a Phone Number

python scripts/telnyx_api.py get-available-phone                          # Get first available
python scripts/telnyx_api.py get-connection-id <assistant_id> telephony   # Get connection ID
python scripts/telnyx_api.py assign-phone <phone_number_id> <connection_id> voice  # Assign

High-Level Alternative: Setup Agent in One Step

python scripts/telnyx_api.py setup-agent "find-window-washing-contractors" "Contractor Caller" "You are calling to get quotes for commercial window washing. Ask about: rates per floor, availability, insurance. Be professional." "Hi, I am calling to inquire about your commercial window washing services. Do you have a moment to discuss rates?"

This automatically creates the assistant, links it to the mission run, finds an available phone number, assigns it, and saves all IDs to the state file.

⚠️ The slug MUST match what init created. If it doesn't, the agent won't be linked and scheduled events won't appear on the frontend.

Verify linking worked immediately after:

python scripts/telnyx_api.py list-linked-agents <mission_id> <run_id>
# Must show your assistant_id. If empty → slug was wrong. Fix with:
python scripts/telnyx_api.py link-agent <mission_id> <run_id> <assistant_id>

Step 2.3: Link Agent to Mission Run

If using setup-agent: Linking is done automatically (only if slug matches init).

If setting up manually:

python scripts/telnyx_api.py link-agent <mission_id> <run_id> <assistant_id>
python scripts/telnyx_api.py list-linked-agents <mission_id> <run_id>
python scripts/telnyx_api.py unlink-agent <mission_id> <run_id> <assistant_id>

Phase 3: Scheduling Calls/SMS

Business Hours Consideration

CRITICAL: Before scheduling calls, consider business hours (9 AM - 5 PM local time). scheduled_at must be in the future (at least 1 minute from now).

python scripts/telnyx_api.py schedule-call <assistant_id> "+15551234567" "+15559876543" "2024-12-01T14:30:00Z" <mission_id> <run_id>
python scripts/telnyx_api.py schedule-sms <assistant_id> "+15551234567" "+15559876543" "2024-12-01T14:30:00Z" "Your message here"

Save the returned event id.

Phase 4: Monitoring Call Completion

⚠️ THIS IS THE MOST IMPORTANT PART

After scheduling a call or SMS, Telnyx executes it autonomously at the scheduled time. You need to poll to find out when it's done, then update the mission accordingly.

Check Scheduled Event Status

python scripts/telnyx_api.py get-event <assistant_id> <event_id>

Use Cron Jobs to Poll

Use your bot's cron system to schedule polling. Do NOT block the main session waiting. Match the poll interval to the expected wait time:

If you know the exact scheduled time, don't start polling until after that time. Schedule your first poll for scheduled_time + 2 minutes.

Cron Job Pattern

When you schedule a call/SMS, create a cron job to poll for it:

Create cron: poll at appropriate interval
  → Run get-event <assistant_id> <event_id>
  → If completed: update step, log event, complete mission if last step, DELETE THIS CRON
  → If failed: update step as failed, log error, DELETE THIS CRON
  → If pending/in_progress: do nothing, cron runs again at next interval

⚠️ ALWAYS clean up cron jobs when a mission reaches a terminal state (completed, failed, cancelled). Never leave polling crons running after a mission ends.

Adjusting Poll Frequency

You can update the cron interval as circumstances change:

• Scheduled for 2 weeks from now? Start with 8-hour polling.
• As the scheduled time approaches (within 1 hour), tighten to 5-minute intervals.
• After the scheduled time passes, tighten to 30-second intervals.

Event Status Values

Call Status Values (Phone Calls Only)

Phase 5: Getting Conversation Insights

Once a call completes with a conversation_id, retrieve insights. Poll until status is "completed" (wait 10 seconds between retries).

python scripts/telnyx_api.py get-insights <conversation_id>

Telnyx automatically creates default insight templates when an assistant is created. You don't need to manage these — just read the results.

Phase 6: Complete the Mission

python scripts/telnyx_api.py update-run <mission_id> <run_id> succeeded

# Or with full results:
python scripts/telnyx_api.py complete "find-window-washing-contractors" <mission_id> <run_id> "Summary of results" '{"key": "payload"}'

Event Logging Reference

Log EVERY action as an event. Always update step status via update-step AND log corresponding events.

# When STARTING a step:
python scripts/telnyx_api.py update-step "$MISSION_ID" "$RUN_ID" "research" "in_progress"
python scripts/telnyx_api.py log-event "$MISSION_ID" "$RUN_ID" step_started "Starting: Research" "research"

# When COMPLETING a step:
python scripts/telnyx_api.py update-step "$MISSION_ID" "$RUN_ID" "research" "completed"
python scripts/telnyx_api.py log-event "$MISSION_ID" "$RUN_ID" step_completed "Completed: Research" "research"

# When a step FAILS:
python scripts/telnyx_api.py update-step "$MISSION_ID" "$RUN_ID" "calls" "failed"
python scripts/telnyx_api.py log-event "$MISSION_ID" "$RUN_ID" error "Failed: Could not reach contractors" "calls"

Quick Reference: All Python Commands

# Check setup
python scripts/telnyx_api.py check-key

# Missions
python scripts/telnyx_api.py create-mission <name> <instructions>
python scripts/telnyx_api.py get-mission <mission_id>
python scripts/telnyx_api.py list-missions

# Runs
python scripts/telnyx_api.py create-run <mission_id> <input_json>
python scripts/telnyx_api.py get-run <mission_id> <run_id>
python scripts/telnyx_api.py update-run <mission_id> <run_id> <status>
python scripts/telnyx_api.py list-runs <mission_id>

# Plan
python scripts/telnyx_api.py create-plan <mission_id> <run_id> <steps_json>
python scripts/telnyx_api.py get-plan <mission_id> <run_id>
python scripts/telnyx_api.py update-step <mission_id> <run_id> <step_id> <status>

# Events
python scripts/telnyx_api.py log-event <mission_id> <run_id> <type> <summary> <step_id> [payload_json]
python scripts/telnyx_api.py list-events <mission_id> <run_id>

# Assistants
python scripts/telnyx_api.py list-assistants [--name=<filter>] [--page=<n>] [--size=<n>]
python scripts/telnyx_api.py create-assistant <name> <instructions> <greeting> [options_json]
python scripts/telnyx_api.py get-assistant <assistant_id>
python scripts/telnyx_api.py update-assistant <assistant_id> <updates_json>
python scripts/telnyx_api.py get-connection-id <assistant_id> [telephony|messaging]

# Phone Numbers
python scripts/telnyx_api.py list-phones [--available]
python scripts/telnyx_api.py get-available-phone
python scripts/telnyx_api.py assign-phone <phone_id> <connection_id> [voice|sms]

# Scheduled Events
python scripts/telnyx_api.py schedule-call <assistant_id> <to> <from> <datetime> <mission_id> <run_id>
python scripts/telnyx_api.py schedule-sms <assistant_id> <to> <from> <datetime> <text>
python scripts/telnyx_api.py get-event <assistant_id> <event_id>
python scripts/telnyx_api.py cancel-scheduled-event <assistant_id> <event_id>
python scripts/telnyx_api.py list-events-assistant <assistant_id>

# Insights
python scripts/telnyx_api.py get-insights <conversation_id>

# Mission Run Agents
python scripts/telnyx_api.py link-agent <mission_id> <run_id> <telnyx_agent_id>
python scripts/telnyx_api.py list-linked-agents <mission_id> <run_id>
python scripts/telnyx_api.py unlink-agent <mission_id> <run_id> <telnyx_agent_id>

# State Management
python scripts/telnyx_api.py list-state
python scripts/telnyx_api.py get-state <slug>
python scripts/telnyx_api.py remove-state <slug>

# Memory
python scripts/telnyx_api.py save-memory <slug> <key> <value_json>
python scripts/telnyx_api.py get-memory <slug> [key]
python scripts/telnyx_api.py append-memory <slug> <key> <item_json>

# High-Level Workflows
python scripts/telnyx_api.py init <name> <instructions> <request> [steps_json]
python scripts/telnyx_api.py setup-agent <slug> <name> <instructions> <greeting>
python scripts/telnyx_api.py complete <slug> <mission_id> <run_id> <summary> [payload_json]

Complete Example: SMS Mission with Cron Polling

Here's the full flow for an SMS mission with proper lifecycle tracking and cron-based polling:

# 1. Init mission
python scripts/telnyx_api.py init "SMS Test 003" \
  "Send a test SMS to +13322200013" \
  "SMS test with full tracking" \
  '[{"step_id": "setup", "description": "Create SMS agent", "sequence": 1},
    {"step_id": "sms", "description": "Schedule SMS", "sequence": 2},
    {"step_id": "verify", "description": "Verify delivery", "sequence": 3}]'
# Save: mission_id, run_id

# 2. Step 1: Setup agent
python scripts/telnyx_api.py update-step $MISSION_ID $RUN_ID setup in_progress
python scripts/telnyx_api.py log-event $MISSION_ID $RUN_ID step_started "Starting: Create SMS agent" setup

python scripts/telnyx_api.py setup-agent "sms-test-003" "SMS Agent" "Send test messages" "Test from bot"
# Save: assistant_id, phone_number

python scripts/telnyx_api.py update-step $MISSION_ID $RUN_ID setup completed
python scripts/telnyx_api.py log-event $MISSION_ID $RUN_ID step_completed "Completed: Created assistant $ASSISTANT_ID" setup

# 3. Step 2: Schedule SMS
python scripts/telnyx_api.py update-step $MISSION_ID $RUN_ID sms in_progress
python scripts/telnyx_api.py log-event $MISSION_ID $RUN_ID step_started "Starting: Schedule SMS" sms

python scripts/telnyx_api.py schedule-sms $ASSISTANT_ID "+13322200013" "$PHONE" "2026-02-19T18:43:00Z" \
  "What do you call a bear with no teeth? A gummy bear!" \
  $MISSION_ID $RUN_ID sms
# Save: event_id

python scripts/telnyx_api.py update-step $MISSION_ID $RUN_ID sms completed
python scripts/telnyx_api.py log-event $MISSION_ID $RUN_ID step_completed "Completed: SMS scheduled" sms

# 4. Step 3: Verify delivery — CREATE A CRON JOB TO POLL
python scripts/telnyx_api.py update-step $MISSION_ID $RUN_ID verify in_progress
python scripts/telnyx_api.py log-event $MISSION_ID $RUN_ID step_started "Starting: Poll for delivery" verify

# >>> Create a cron job that fires AFTER the scheduled time <<<
# >>> Cron runs: get-event $ASSISTANT_ID $EVENT_ID <<<
# >>> On completed: update-step verify completed, log-event, update-run succeeded, DELETE CRON <<<
# >>> On failed: update-step verify failed, log-event, update-run failed, DELETE CRON <<<
# >>> On pending/in_progress: do nothing, cron fires again next interval <<<

# 5. (Cron fires, detects completion)
python scripts/telnyx_api.py update-step $MISSION_ID $RUN_ID verify completed
python scripts/telnyx_api.py log-event $MISSION_ID $RUN_ID step_completed "Completed: SMS delivered" verify
python scripts/telnyx_api.py update-run $MISSION_ID $RUN_ID succeeded
# DELETE the polling cron job!

Mission Classes

Not all missions are the same. Identify which class before planning.

Does call N depend on results of call N-1?
  YES -> Is it negotiation (leveraging previous results)?
    YES -> Class 3: Sequential Negotiation
    NO  -> Does it have distinct rounds with human approval?
      YES -> Class 4: Multi-Round / Follow-up
      NO  -> Class 5: Information Gathering -> Action
  NO  -> Do you need structured scoring/ranking?
    YES -> Class 2: Parallel Screening with Rubric
    NO  -> Class 1: Parallel Sweep

Class 1: Parallel Sweep

Fan out calls in parallel batches. Same question to many targets. Schedule all calls in one batch (stagger by 1-2 min). Analysis happens after ALL calls complete.

Class 2: Parallel Screening with Rubric

Fan out calls in parallel with structured scoring criteria. Results are ranked post-hoc via insights.

Class 3: Sequential Negotiation

Calls MUST run serially. Each call's strategy depends on previous results. Use update-assistant between calls to inject context. Never parallelize these.

Class 4: Multi-Round / Follow-up

Two or more distinct phases. Round 1 is broad outreach, human approval gate, then Round 2 targets a subset.

Class 5: Information Gathering -> Action

Call to find something, then act on it. Early termination when goal is met — cancel remaining calls.

Operational Guide

Default Tools

The send_dtmf tool is included by default. Most outbound calls hit an IVR first.

IVR Navigation

Expect IVRs even when calling businesses. Instruct the assistant to press 0 or say 'representative'.

Call Limits and Throttling

Stagger calls in batches of 5-10, space scheduled times 1-2 minutes apart, monitor for 429 errors.

Answering Machine Detection (AMD)

• Enable for human contacts (leave voicemail or skip machines)
• Disable for IVR systems, businesses with phone trees — set action to continue_assistant

Polling for Results: Use Cron Jobs

After scheduling calls, set up a cron job to poll periodically. Don't block the main session.

Retry Strategy

Track every number's status in mission memory. Retry based on recipient type:

• Automated systems: retry in 5-15 min, up to 3 times
• Service industry: retry in 30 min - 2 hours, avoid peak hours
• Professionals: retry next business day, leave one voicemail max

Approval Requests (Sensitive Actions)

For destructive or sensitive actions during voice calls, request user approval first:

./scripts/approval.sh request "Delete GitHub repo myproject"
./scripts/approval.sh request "Send $500 to John" --biometric
./scripts/approval.sh request "Post tweet about X" --details "Full text: ..."

When to request approval:

• Deleting repos, files, or data
• Sending money or making purchases
• Posting to social media
• Sending emails/messages to others
• Any irreversible action

Response values:

• approved → Execute the action, confirm completion
• denied → Tell user "Okay, I won't do that"
• timeout → "I didn't get a response, should I try again?"
• no_devices → Skip approval, action not executed (no mobile app)

Example flow in voice call:

1. User: "Delete my test-repo on GitHub"
2. You: "I'll need your approval for that. Check your phone."
3. Run: approval.sh request "Delete GitHub repo test-repo"
4. If approved: Delete the repo, then say "Done, test-repo has been deleted"
5. If denied: "Got it, I won't delete it"

Gateway Requirements

Voice calls route requests to the main agent via sessions_send. This tool is blocked by default on the Gateway HTTP tools API. You must explicitly allow it:

// In openclaw.json → gateway.tools
{
  "gateway": {
    "tools": {
      "allow": ["sessions_send"]
    }
  }
}

Or via CLI: openclaw config patch '{"gateway":{"tools":{"allow":["sessions_send"]}}}'

Without this, voice calls will connect but the agent won't be able to process any requests (deep tool calls return 404).

⚠️ WARNING: This MUST go under gateway.tools.allow, NOT top-level tools.allow. The top-level tools.allow is the agent's tool allowlist — putting sessions_send there will restrict your agent to ONLY that tool, breaking everything. If you accidentally did this, remove the top-level tools.allow entry and restart.

❌ Common Pitfalls

Configuration

Edit skill-config.json:

Environment variables for the Python missions API:

• CLAWDTALK_API_KEY — your ClawdTalk API key (required for missions)
• CLAWDTALK_API_URL — override the API endpoint (default: https://clawdtalk.com/v1)

Troubleshooting

• Auth failed: Regenerate API key at clawdtalk.com
• Gateway token/port changed: Re-run ./setup.sh to update skill-config.json with the new values
• Empty responses: Run ./setup.sh and restart gateway
• Slow responses: Try a faster model in your gateway config
• Debug mode: DEBUG=1 ./scripts/connect.sh restart
• Missions API key: Run python scripts/telnyx_api.py check-key to verify
• JSON parsing errors: Use single quotes around JSON arguments