"Operate OpenSquilla messaging channels, pairings, certification,
Scanned 9/8/2026
Install to Claude Code
npx -y skills add VectorSpaceLab/AREX-Skill --skill channels-and-integrations --agent claude-codeInstalls into .claude/skills of the current project.
Are you the author of Channels And Integrations?
Add the live security badge to your README — it updates automatically with every re-scan.
[](https://www.skillsdirectory.com/skills/vectorspacelab-channels-and-integrations)More formats (shields.io, HTML) on the badges page.
---
name: channels-and-integrations
description: "Operate OpenSquilla messaging channels, pairings, certification,
adapter status, and the MCP server bridge."
metadata:
disco-role: operating
disable-model-invocation: true
license: Apache 2.0
---
# Channels and Integrations
Use this sub-skill when the task is about OpenSquilla's ingress/egress integration layer on top of the gateway:
- messaging channel catalog, setup, CRUD, status, restart, logout, and live-vs-offline validation;
- channel admission, direct-message pairing, allowlists, group-session scope, approval-code delivery, and operator pairings;
- adapter capability/profile evidence, maturity, delivery diagnostics, transport leases, and certification probes;
- the stdio MCP server bridge launched by `opensquilla mcp-server run`.
This guidance targets OpenSquilla 0.5.3 behavior. Installed-package evidence for this surface included working OpenSquilla help, the channel catalog, and a successful `mcp` import; live provider credentials were not assumed.
## Route before acting
- Gateway installation, gateway lifecycle, bind/port, public exposure, Web UI basics, and `doctor` readiness checks: use [`../setup-and-gateway/SKILL.md`](../setup-and-gateway/SKILL.md).
- Provider API keys, model/router tiers, search providers, and general config precedence: use [`../configuration-and-routing/SKILL.md`](../configuration-and-routing/SKILL.md).
- Non-interactive `agent`/`chat` runs, sessions, diagnostics, sandbox commands, replay, cron, bundle, and automation flags: use [`../cli-and-automation/SKILL.md`](../cli-and-automation/SKILL.md).
- Terminal UI, desktop shell, Web UI presentation, or surface-specific UI failures: use [`../tui-and-desktop/SKILL.md`](../tui-and-desktop/SKILL.md).
Stay in this sub-skill for the integration boundary itself: a channel or MCP bridge may trigger agent turns, tool calls, approvals, and session updates, but those surfaces are reached through the same gateway-backed runtime.
## Fast operating map
1. **Discover the channel catalog.** Run `opensquilla channels types --json` and `opensquilla channels describe <type> --json`; do not assume provider fields from memory. See [`references/commands.md`](references/commands.md).
2. **Add or edit channel config.** Use `opensquilla configure channels` for interactive setup, or `opensquilla channels add|edit` for scripted changes. Config edits require a gateway process restart, not just `channels restart <name>`. See [`references/channel-workflows.md`](references/channel-workflows.md).
3. **Check runtime, not just saved config.** `opensquilla channels list` proves config exists; `opensquilla channels status <name> --json` proves whether the running gateway loaded it and exposes status/capability/diagnostic data. See [`references/runtime-contracts.md`](references/runtime-contracts.md).
4. **Handle pairings deliberately.** Unknown authenticated DMs default to pairing mode. Use `opensquilla channels pairings list|approve|revoke` and treat `--admin` as a host-trust decision for that channel sender. See [`references/runtime-contracts.md`](references/runtime-contracts.md).
5. **Separate offline validation from live certification.** Local schema validation and status are not end-to-end provider proof. `opensquilla channels certify` is environment-only and redacts evidence; outbound tests require explicit side-effect flags and targets. See [`references/channel-workflows.md`](references/channel-workflows.md).
6. **Run the MCP bridge as stdio.** Start a gateway first, then configure the MCP-capable client to launch `opensquilla mcp-server run` or `opensquilla mcp-server run --gateway ws://host:port/ws`. See [`references/mcp-server-bridge.md`](references/mcp-server-bridge.md).
7. **Troubleshoot from the boundary outward.** Start with config/schema, gateway restart state, runtime status, pairing/admission diagnostics, delivery ledger state, then provider credentials/network. See [`references/troubleshooting.md`](references/troubleshooting.md).
## Channel catalog scope
The public channel catalog for this build has eight families: `dingtalk`, `discord`, `feishu`, `matrix`, `qq`, `slack`, `telegram`, and `wecom`. Public vendor adapters are experimental; catalog presence means the adapter is registered and has a setup schema, not that every provider feature is certified.
Important catalog rules:
- `channels describe <type>` is the source of truth for current required fields, secret fields, mode choices, public-URL requirements, optional extras, and restart behavior.
- `matrix` declares the `matrix` optional extra. The MCP bridge separately requires the `mcp` extra.
- Microsoft Teams code exists only as a hidden legacy adapter in this version; do not present it as a supported public channel.
## Safe defaults
- Keep gateway-bound integration surfaces on loopback unless a webhook provider explicitly requires public reachability and the operator has configured a trusted reverse proxy/tunnel plus authentication.
- Prefer `dm_access=pairing` for DMs. Use `open` only for intentionally public bots; use `allowlist` when provider sender IDs are known.
- Keep group sessions in the default per-sender scope unless the room is intentionally collaborative.
- Use the narrowest OpenSquilla permissions and sandbox posture that can complete the task; channel approval or pairing does not bypass host-side tool policy.
- Never paste provider keys, bot tokens, signing secrets, or channel secrets into examples, issue text, source, or generated skill files.
## Bundled scripts
No runtime helper script is bundled for this sub-skill. The useful operations are already exposed as safe OpenSquilla CLI commands, while the repo-owned live smoke scripts require provider credentials, network side effects, or maintainer harness context and are reference-only or excluded from runtime use.
Is this your skill, or is something wrong with this listing? Request removal or report an issue. Author removals are honored within 72 hours.
No comments yet. Be the first to comment!