# Send Openclaw phone to your agent
Hand the extracted package to your coding agent with a concrete install brief instead of figuring it out manually.
## Fast path
- Download the package from Yavira.
- Extract it into a folder your agent can access.
- Paste one of the prompts below and point your agent at the extracted folder.
## Suggested prompts
### New install

```text
I downloaded a skill package from Yavira. Read SKILL.md from the extracted folder and install it by following the included instructions. Then review README.md for any prerequisites, environment setup, or post-install checks. Tell me what you changed and call out any manual steps you could not complete.
```
### Upgrade existing

```text
I downloaded an updated skill package from Yavira. Read SKILL.md from the extracted folder, compare it with my current installation, and upgrade it while preserving any custom configuration unless the package docs explicitly say otherwise. Then review README.md for any prerequisites, environment setup, or post-install checks. Summarize what changed and any follow-up checks I should run.
```
## Machine-readable fields
```json
{
  "schemaVersion": "1.0",
  "item": {
    "slug": "openclaw-phone",
    "name": "Openclaw phone",
    "source": "tencent",
    "type": "skill",
    "category": "开发工具",
    "sourceUrl": "https://clawhub.ai/BenjaminWaye/openclaw-phone",
    "canonicalUrl": "https://clawhub.ai/BenjaminWaye/openclaw-phone",
    "targetPlatform": "OpenClaw"
  },
  "install": {
    "downloadUrl": "/downloads/openclaw-phone",
    "sourceDownloadUrl": "https://wry-manatee-359.convex.site/api/v1/download?slug=openclaw-phone",
    "sourcePlatform": "tencent",
    "targetPlatform": "OpenClaw",
    "packageFormat": "ZIP package",
    "primaryDoc": "SKILL.md",
    "includedAssets": [
      "README.md",
      "SKILL.md",
      "examples/prompts.md",
      "docs/auth-config.md",
      "docs/event-updates.md",
      "agents/openai.yaml"
    ],
    "downloadMode": "redirect",
    "sourceHealth": {
      "source": "tencent",
      "slug": "openclaw-phone",
      "status": "healthy",
      "reason": "direct_download_ok",
      "recommendedAction": "download",
      "checkedAt": "2026-05-06T16:53:13.762Z",
      "expiresAt": "2026-05-13T16:53:13.762Z",
      "httpStatus": 200,
      "finalUrl": "https://wry-manatee-359.convex.site/api/v1/download?slug=openclaw-phone",
      "contentType": "application/zip",
      "probeMethod": "head",
      "details": {
        "probeUrl": "https://wry-manatee-359.convex.site/api/v1/download?slug=openclaw-phone",
        "contentDisposition": "attachment; filename=\"openclaw-phone-1.0.1.zip\"",
        "redirectLocation": null,
        "bodySnippet": null,
        "slug": "openclaw-phone"
      },
      "scope": "item",
      "summary": "Item download looks usable.",
      "detail": "Yavira can redirect you to the upstream package for this item.",
      "primaryActionLabel": "Download for OpenClaw",
      "primaryActionHref": "/downloads/openclaw-phone"
    },
    "validation": {
      "installChecklist": [
        "Use the Yavira download entry.",
        "Review SKILL.md after the package is downloaded.",
        "Confirm the extracted package contains the expected setup assets."
      ],
      "postInstallChecks": [
        "Confirm the extracted package includes the expected docs or setup files.",
        "Validate the skill or prompts are available in your target agent workspace.",
        "Capture any manual follow-up steps the agent could not complete."
      ]
    }
  },
  "links": {
    "detailUrl": "https://openagent3.xyz/skills/openclaw-phone",
    "downloadUrl": "https://openagent3.xyz/downloads/openclaw-phone",
    "agentUrl": "https://openagent3.xyz/skills/openclaw-phone/agent",
    "manifestUrl": "https://openagent3.xyz/skills/openclaw-phone/agent.json",
    "briefUrl": "https://openagent3.xyz/skills/openclaw-phone/agent.md"
  }
}
```
## Documentation

### CallMyCall (OpenClaw Skill)

This skill helps you operate CallMyCall from chat. It is pull based (no webhook callbacks): you start a call, store the returned call ID in state, and later fetch status and results on demand.

### API Key Resolution (OpenClaw best practice)

Resolve credentials in this order:

Environment variable: CALLMYCALL_API_KEY (preferred)
OpenClaw user config: ~/.openclaw/openclaw.json under skills.openclaw-phone.apiKey
If still missing, ask for a one-time key for the current task only.
Only if the user explicitly asks for persistence, provide manual instructions for saving the key to ~/.openclaw/openclaw.json.

Persistence rules:

Never store API keys in SKILL.md, examples, references, or memory/state files.
Do not write API keys into recent_calls or any conversation-visible output. Do not tell the user “I won’t echo it back.”
Use interactive keys for the current task only.
Do not write user config files automatically from this skill.

### How This Skill Should Work

Resolve API key using the order in "API Key Resolution (OpenClaw best practice)".
Collect required call info using a layered gating flow (below).
Present a short review summary and ask for confirmation.
On confirmation, call POST /v1/start-call.
Store the returned sid in state as a recent call.

### Required Auth

Send the API key in the Authorization header:

Authorization: Bearer YOUR_API_KEY

Never echo the API key back to the user or include it in logs/summaries.

### Stateful Calls List (required)

Maintain a list (last 10) of recent calls in state:

recent_calls: array of objects

id (call SID)
phone_number
task
started_at
status (optional, updated when you fetch)

Use this list to let the user say "end call 1" or "show results for call 2".

### Layered Gating Flow (copy from web app)

Do not rely on a single validation step. Use all layers below.

### Layer 1: Structured collection contract

Do not finalize a task until all required fields exist:

phone_number
language
call_brief (background + goals)

### Layer 2: Task gap analysis

When the user gives the initial request, analyze what is missing. Then ask only for missing info. If the user answers partially, repeat analysis and keep asking for the remaining gaps.

### Layer 3: Prompt level enforcement

While missing info exists, continue gathering required fields. Do not proceed to the call until all required fields are present.

### Layer 4: Runtime validation before finalizing

Before sending the call request:

Ensure phone exists and is E.164
Block emergency or premium numbers
Ensure from_number is not the same as phone_number
If from_number is requested, run caller-ID preflight:

GET /v1/verified-caller-ids
Confirm requested from_number exists in verified_caller_ids
If not verified: do not place call yet; ask user whether to continue with default caller ID or verify first


Normalize language; normalize voice fields (genderVoice, openaiVoice) only if provided
If scheduling is present, parse and clamp to a valid time

### Layer 5: Human review gate

Present a short review summary:

Phone number
Call brief (background + goals)
Language (and voice if provided)
Any schedule

Ask: "Confirm and place the call?" Do not proceed without explicit confirmation.

### Start a Call

Collect required fields using the layered gating flow.
If from_number is provided, run caller-ID preflight via GET /v1/verified-caller-ids.
If requested from_number is not verified, ask user to choose:

continue now with default caller ID, or
verify number first (POST /v1/verify-caller-id, then GET /v1/verification-status/:verificationId).


If a schedule/time is requested, follow Scheduled Requests (No Cron) below instead of calling the API immediately.
Otherwise call POST /v1/start-call.
Store the returned sid in recent_calls.
Reply with confirmation and the call ID.

### Scheduled Requests (No Cron)

Because the API has no scheduling field:

Collect all required fields now.
Save a compact call plan in skill state only for in-session follow-up.
Do not create or modify OS schedulers (cron, launchd, task scheduler) and do not run autonomous background turns.
Offer one of these safe options:

place the call now, or
provide a reminder summary and ask the user to return at the target time to run start-call.

If the user asks to schedule for later, explain that this skill does not create background jobs; it can prepare the call plan and execute when the user confirms in-session.

### List Recent Calls

Read recent_calls from state.
For each call, fetch status via GET /v1/calls/:callId if needed.
Display a numbered list.

### Retry Until Answered (important)

When the user asks to call repeatedly until answered:

Place one call with POST /v1/start-call.
Poll GET /v1/calls/:callId until terminal status.
Treat response as either flat (status, duration) or nested (call.status, call.duration).
If status is busy, no-answer, failed, or canceled, wait requested interval and place next call.
Stop retry loop when:

status is in-progress, or
status is completed with duration > 0.


Report each attempt (call ID + status) back to user.

Implementation note: keep one base URL per run (https://call-my-call-backend.fly.dev preferred) and use it consistently for both start + status endpoints.

### End a Call

If the user says "end the call" without specifying which, list recent calls and ask which one.

If there is only one active call, confirm and end it.

Call:

POST /v1/end-call with { callSid }.

### Get Results

When the user asks for call results:

Fetch status via GET /v1/calls/:callId.
If available, fetch transcript via GET /v1/calls/:callId/transcripts/stream.
If the call was recorded, fetch recording URL via GET /v1/calls/:callSid/recording.

Return:

Status (completed, failed, canceled)
Short summary (1 to 3 bullets)
Transcript excerpt (first few lines, only after user asks to view transcript content)
Recording URL (if present, warn that URL access may expose sensitive audio)

### Safety and UX

If user input is ambiguous, ask a clarification question.
Never expose secrets or store API keys in transcript.
Treat transcripts and recordings as sensitive; share only minimal excerpts requested by the user.
Never create persistent scheduler entries or autonomous background execution from this skill.
If a request fails, show the HTTP error and suggest next steps.

### References

Full API reference: references/api.md
Examples: examples/prompts.md
## Trust
- Source: tencent
- Verification: Indexed source record
- Publisher: BenjaminWaye
- Version: 1.0.1
## Source health
- Status: healthy
- Item download looks usable.
- Yavira can redirect you to the upstream package for this item.
- Health scope: item
- Reason: direct_download_ok
- Checked at: 2026-05-06T16:53:13.762Z
- Expires at: 2026-05-13T16:53:13.762Z
- Recommended action: Download for OpenClaw
## Links
- [Detail page](https://openagent3.xyz/skills/openclaw-phone)
- [Send to Agent page](https://openagent3.xyz/skills/openclaw-phone/agent)
- [JSON manifest](https://openagent3.xyz/skills/openclaw-phone/agent.json)
- [Markdown brief](https://openagent3.xyz/skills/openclaw-phone/agent.md)
- [Download page](https://openagent3.xyz/downloads/openclaw-phone)