concinno-skills-mobile-ext

Extended enterprise-messaging agent skills for Concinno. Complements the sibling concinno-skills-mobile (Twilio SMS / Voice / WhatsApp) by adding three channels Twilio does not natively cover: Microsoft Teams, WhatsApp Business Cloud API (direct Meta, no Twilio), and WeCom (企业微信).

All three tools are thin httpx REST wrappers — zero new runtime dependencies beyond Concinno itself.

Status

MVP (0.1.0) — three one-shot tools:

Install

pip install concinno-skills-mobile-ext

Credentials

Each platform's credentials live under well-known keys in the Concinno CredentialStore, which reads (in order):

1. Process runtime overrides via CredentialStore.set(...).
2. Env var CONCINNO_CRED_<UPPER_KEY>.
3. ~/.concinno/credentials.json.

Microsoft Teams

Bring your own Azure AD OAuth Bearer token (e.g. produced via msal or a manual consent flow). This package never opens a browser.

WhatsApp Business Cloud API (Meta)

Both values come from the Meta Business Suite / App dashboard.

WeCom (企业微信)

Two supported shapes — group-bot webhook (simplest) is tried first:

…or app-mode credentials:

Example ~/.concinno/credentials.json:

{
  "teams_access_token": "eyJ0eXAiOi…",
  "whatsapp_access_token": "EAAB…",
  "whatsapp_phone_number_id": "1234567890",
  "wecom_webhook_url": "https://qyapi.weixin.qq.com/cgi-bin/webhook/send?key=abcd-1234"
}

If a required token is missing the tool returns {"error": "no <service> credentials — set via CredentialStore / env"} rather than crashing.

Usage via Concinno ToolRegistry

When the consumer sets CONCINNO_LOAD_PLUGINS=1, the default registry auto-mounts every tool:

import os
os.environ["CONCINNO_LOAD_PLUGINS"] = "1"

from concinno.tools.registry import get_default_registry

reg = get_default_registry()
assert {"TeamsSend", "WhatsAppOfficial", "WeComSend"} <= set(
    reg.list_deferred()
)

teams = reg.get("TeamsSend")
teams.call(
    action="send",
    chat_id="19:2da4c29f6d7041eca70b638b43d45437@thread.v2",
    content="Deploy succeeded",
    content_type="text",
)

Direct Python usage

from concinno_skills_mobile_ext import (
    TeamsSend,
    WhatsAppOfficial,
    WeComSend,
)

TeamsSend().call(
    action="send",
    chat_id="19:abc@thread.v2",
    content="<b>Deploy</b> succeeded",
    content_type="html",
)

WhatsAppOfficial().call(
    action="send",
    to="+15551234567",
    body="Your verification code is 123456",
)

WeComSend().call(
    action="send_webhook",
    content="CI pipeline green on main",
    mentioned_list=["@all"],
)

WeComSend().call(
    action="send_app",
    content="Private alert",
    user_ids=["alice", "bob"],
)

All tools return {"ok": True, ...} or a structured payload on success, or {"error": "..."} on failure — same shape as the other Concinno built-in tools. No exceptions escape the call() surface.

Concurrency

All three tools set is_concurrency_safe = False:

• Teams — per-app/user Graph throttling and conversational ordering.
• WhatsApp — per-phone-number Meta rate limits + 24-hour conversation window semantics.
• WeCom — shared access-token cache in app mode; per-bot rate limits in webhook mode.

The Concinno scheduler honours this automatically.

Safety caps

• Teams content ≤ 28 000 chars (text or HTML).
• WhatsApp body ≤ 4 096 chars (Meta text message cap).
• WeCom content ≤ 2 048 bytes UTF-8 (WeCom cap; CJK text is counted at 3 bytes/char).

API pins

• Microsoft Graph: v1.0 (POST /chats/{id}/messages).
• Meta WhatsApp Cloud API: v21.0 (POST /{ver}/{phone_number_id}/messages).
• WeCom:
  • App — POST /cgi-bin/message/send?access_token=… (access token obtained via GET /cgi-bin/gettoken, cached 7200 s with a 60 s refresh slack).
  • Webhook — POST /cgi-bin/webhook/send?key=….

Any version bump must call out the full regression-test matrix in RELEASE_COORDINATION.md.

License

Apache-2.0. See LICENSE in the Concinno monorepo.