Implement authenticated chat ingress (codex)

[numnx]

🧩 Task: Implement authenticated chat ingress

Automated task PR opened by Code UX for T04.

📋 Summary

Sprint: Sprint 162 · Integrate Chat Connections · Task: T04

Integrate configurable external chat providers under Settings -> Providers so WhatsApp, iMessage, Telegram, Slack, Microsoft Teams, and Discord can be securely connected to project-manager chat flows, managed through dashboard and MCP APIs, linked flexibly across projects/channels, and routed through the existing chat reply runtime with external-channel-safe replies that suppress dashboard-only rich widgets.

📝 Full Task Prompt

## Objective
Implement authenticated inbound chat-provider ingress that verifies provider bridge requests, deduplicates external messages, resolves channel bindings to projects, and posts external user instructions into existing conversation threads.

## Scope
- Create `src/services/chat-provider-ingress-service.ts`
- Create `src/services/chat-provider-security.ts`
- Create `src/server/chat-provider-ingress-routes.ts`
- Edit `src/server/dashboard-server.ts`
- Edit `src/server/dashboard-route-registration.ts`
- Edit `src/app/dependency-factory/dashboard-factory.ts`
- Edit `src/app/lifecycle/dashboard-lifecycle-service.ts`
- Add `tests/backend/services/chat-provider-ingress-service.test.ts`
- Add `tests/backend/server/chat-provider-ingress-routes.test.ts`
- Consumes repository and route wiring from T01 and T02

## Implementation Requirements
1. Add inbound endpoints for OpenClaw/webhook/native bridge POST payloads, including `POST /api/chat-providers/ingress/:providerConnectionId` and a provider-kind-compatible payload parser for WhatsApp, iMessage, Telegram, Slack, Microsoft Teams, and Discord fields.
2. Verify each inbound request with the configured bridge secret: bearer token for OpenClaw/native bridges, HMAC signature for webhook mode when configured, timestamp freshness, replay prevention, and per-provider connection enablement.
3. Normalize inbound payloads into one internal message shape containing provider connection id, provider kind, external channel id/name, external sender id/name, text body, external message id, timestamp, and raw metadata with secrets stripped.
4. Use repository idempotency checks before posting messages; duplicates must return the existing delivery/conversation result without creating another conversation message.
5. Resolve enabled channel bindings by provider connection and channel id; if multiple projects match, apply configured routing hints or project selector prefixes first, otherwise record a disambiguation-needed delivery state and return a clear response that does not guess a project.
6. Post routed inbound text through `ChatThreadRuntimeService.postMessage` with metadata marking `source: "chat_provider"`, provider kind, external channel id, external sender, inbound delivery id, and `suppressRichWidgets: true`.
7. Add structured, redacted logs with correlation ids for accepted, rejected, duplicate, ambiguous, and failed ingress attempts.

## Constraints
- Do not implement outbound provider replies or widget stripping in this task.
- Do not accept unauthenticated inbound requests except in explicit test-only paths.
- Do not silently route ambiguous shared channels to the wrong project.
- Do not log raw secrets, bearer tokens, signatures, or full raw provider payloads containing credentials.

## Verification
- Run `pnpm exec vitest run tests/backend/services/chat-provider-ingress-service.test.ts tests/backend/server/chat-provider-ingress-routes.test.ts`.
- Run `pnpm run lint`.
- Success means authenticated inbound messages are normalized, deduplicated, safely routed to project conversation threads, and ambiguous channels are handled without guessing.

✅ QA Review

⏳ QA review runs after this PR is opened. Check the task card in the dashboard for the outcome once it lands.

🤖 Model & Provider

Provider	Codex CLI
Model	gpt-5.5

⏱️ Timing

Started	2026-07-07 06:27:11 UTC
Finished	2026-07-07 06:40:30 UTC
Duration	13m 19s

📊 Token Usage

Metric	Value
Input tokens	145,884
Cached input tokens	5,960,064
Output tokens	30,397
Total tokens	6,136,345
Invocations	1
Tool calls	82
Active time	12m 40s

Included usage estimate (subscription/local login): $4.62

🌿 Branch Info

Base: feature/CODUX-162-integrate-chat-connections
Head: task/feature-codux-162-integrate-chat-connections-t04-codex-mra9o1a6

---

🤖 Generated by Code UX