feat: async resume extraction and CRM apply flow

.env.example

@@ -36,23 +36,28 @@ MINIO_HOST_BIND=127.0.0.1
 MINIO_API_PORT=9000
 MINIO_CONSOLE_PORT=9001
 
-# API (optional defaults; WEBHOOK_SHARED_SECRET required to accept ingest calls)
+# API (optional defaults; API_SHARED_SECRET required to accept ingest calls)
 WEBHOOK_INGEST_HOST=0.0.0.0
 WEBHOOK_INGEST_PORT=8090
 # Required: ingest requests are rejected when unset
-WEBHOOK_SHARED_SECRET=
+API_SHARED_SECRET=
 
 # Worker / consumer (optional defaults)
 WORKER_NAME=integrations-worker
 WORKER_QUEUE_NAMES=jobs.default
 WORKER_BURST=false
+CRM_LINKEDIN_FIELD=cLinkedInUrl
 MAX_ATTACHMENTS_PER_CONTACT=3
 MAX_FILE_SIZE_MB=10
 ALLOWED_FILE_TYPES=pdf,doc,docx,txt
 RESUME_KEYWORDS=resume,cv,curriculum
 OPENAI_API_KEY=
+# For OpenRouter, set OPENAI_BASE_URL=https://openrouter.ai/api/v1
 OPENAI_BASE_URL=
 OPENAI_MODEL=gpt-4o-mini
+# Resume model name without provider prefix; OpenRouter is auto-prefixed to openai/<model>
+RESUME_AI_MODEL=gpt-4o-mini
+RESUME_EXTRACTOR_VERSION=v1
 CRM_SYNC_ENABLED=true
 CRM_SYNC_INTERVAL_SECONDS=900
 CRM_SYNC_PAGE_SIZE=200
@@ -62,6 +67,7 @@ DISCORD_BOT_TOKEN=your_bot_token_here
 HEALTHCHECK_PORT=3000
 DISCORD_SENDMSG_CHARACTER_LIMIT=2000
 CHECK_EMAIL_WAIT=2
+WORKER_API_BASE_URL=http://worker-api:8090
 AUDIT_API_BASE_URL=
 AUDIT_API_TIMEOUT_SECONDS=2.0
 # Required for Discord bot commands that post to channels

.pre-commit-config.yaml

@@ -8,18 +8,11 @@ repos:
       # Run the formatter on changed files only
       - id: ruff-format
 
-  - repo: https://github.com/pre-commit/mirrors-mypy
-    rev: v1.14.0
+  - repo: local
     hooks:
       - id: mypy
-        args: [--strict]
+        name: mypy
+        entry: ./scripts/mypy.sh
+        language: system
+        pass_filenames: false
         files: ^(apps/discord_bot/src/five08/|apps/worker/src/five08/|packages/shared/src/five08/)
-        additional_dependencies:
-          - aiohttp>=3.13.1
-          - discord.py~=2.6.0
-          - pydantic~=2.10
-          - pydantic-settings~=2.8
-          - redis>=6.4.0
-          - requests~=2.31
-          - rq>=2.6.0
-          - types-requests>=2.32.4.20250913

AGENTS.md

@@ -93,7 +93,7 @@ async def setup(bot: commands.Bot) -> None:
 - Add shared env/config in `packages/shared/src/five08/settings.py`.
 - Add service-specific settings in local service `config.py` by subclassing shared settings.
 - Keep secrets in env vars, not code.
-- For Discord CRM audit writes, use `AUDIT_API_BASE_URL` and shared `WEBHOOK_SHARED_SECRET`.
+- For Discord CRM audit writes, use `AUDIT_API_BASE_URL` and shared `API_SHARED_SECRET`.
 
 ## Agent Guidelines
 

DEVELOPMENT.md

@@ -123,9 +123,10 @@ Use `.env.example` as source of truth. Key categories:
 
 - Shared queue/runtime: `REDIS_URL`, `REDIS_QUEUE_NAME`, `POSTGRES_URL`, `JOB_MAX_ATTEMPTS`, `JOB_RETRY_BASE_SECONDS`, `JOB_RETRY_MAX_SECONDS`, `LOG_LEVEL`, webhook settings
 - Bot credentials/integrations: Discord, email, Espo, Kimai
-- Discord CRM audit writer: `AUDIT_API_BASE_URL`, `AUDIT_API_TIMEOUT_SECONDS` (plus shared `WEBHOOK_SHARED_SECRET`)
+- Discord CRM audit writer: `AUDIT_API_BASE_URL`, `AUDIT_API_TIMEOUT_SECONDS` (plus shared `API_SHARED_SECRET`)
 - Worker controls: `WORKER_NAME`, `WORKER_QUEUE_NAMES`, `WORKER_BURST`
-- Worker CRM processing: `MAX_ATTACHMENTS_PER_CONTACT`, `MAX_FILE_SIZE_MB`, `ALLOWED_FILE_TYPES`, `RESUME_KEYWORDS`, `OPENAI_API_KEY`, `OPENAI_BASE_URL`, `OPENAI_MODEL`
+- Worker CRM processing: `MAX_ATTACHMENTS_PER_CONTACT`, `MAX_FILE_SIZE_MB`, `ALLOWED_FILE_TYPES`, `RESUME_KEYWORDS`, `OPENAI_API_KEY`, `OPENAI_BASE_URL`, `OPENAI_MODEL`, `RESUME_EXTRACTOR_VERSION`
+- Resume upload UX wiring: `WORKER_API_BASE_URL` on bot, `CRM_LINKEDIN_FIELD` on worker.
 
 ## CI Notes
 

README.md

@@ -47,6 +47,9 @@ Migrations:
 ### Worker API Endpoints
 
 - `GET /health`: Redis/Postgres/worker health check.
+- `GET /jobs/{job_id}`: Fetch queued job status/result payload.
+- `POST /jobs/resume-extract`: Enqueue resume profile extraction.
+- `POST /jobs/resume-apply`: Enqueue confirmed CRM field apply.
 - `POST /webhooks/{source}`: Generic webhook enqueue endpoint.
 - `POST /webhooks/espocrm`: EspoCRM webhook endpoint (expects array payload).
 - `POST /webhooks/espocrm/people-sync`: EspoCRM contact-change webhook for people cache sync.
@@ -134,7 +137,7 @@ Use `.env.example` as the source of truth for defaults.
 
 ### Worker API Ingest
 
-- `Required` for protected endpoints: `WEBHOOK_SHARED_SECRET` (ingest requests are rejected when unset)
+- `Required` for protected endpoints: `API_SHARED_SECRET` (ingest requests are rejected when unset)
 - `Optional`: `WEBHOOK_INGEST_HOST` (default: `0.0.0.0`)
 - `Optional`: `WEBHOOK_INGEST_PORT` (default: `8090`)
 
@@ -149,18 +152,22 @@ Use `.env.example` as the source of truth for defaults.
 - `Optional`: `CRM_SYNC_ENABLED` (default: `true`)
 - `Optional`: `CRM_SYNC_INTERVAL_SECONDS` (default: `900`)
 - `Optional`: `CRM_SYNC_PAGE_SIZE` (default: `200`)
+- `Optional`: `CRM_LINKEDIN_FIELD` (default: `cLinkedInUrl`)
 - `Optional`: `MAX_ATTACHMENTS_PER_CONTACT` (default: `3`)
 - `Optional`: `MAX_FILE_SIZE_MB` (default: `10`)
 - `Optional`: `ALLOWED_FILE_TYPES` (default: `pdf,doc,docx,txt`)
 - `Optional`: `RESUME_KEYWORDS` (default: `resume,cv,curriculum`)
 - `Optional`: `OPENAI_API_KEY` (if unset, heuristic extraction is used)
-- `Optional`: `OPENAI_BASE_URL`
-- `Optional`: `OPENAI_MODEL` (default: `gpt-4o-mini`)
+- `Optional`: `OPENAI_BASE_URL` (set `https://openrouter.ai/api/v1` for OpenRouter)
+- `Optional`: `RESUME_AI_MODEL` (default: `gpt-4o-mini`; use plain names like `gpt-4o-mini`, OpenRouter gets auto-prefixed to `openai/<model>`)
+- `Optional`: `OPENAI_MODEL` (default: `gpt-4o-mini`; fallback/legacy model setting)
+- `Optional`: `RESUME_EXTRACTOR_VERSION` (default: `v1`; used in resume processing idempotency/ledger keys)
 
 ### Discord Bot Core
 
 - `Required`: `DISCORD_BOT_TOKEN`
 - `Required`: `CHANNEL_ID`
+- `Optional`: `WORKER_API_BASE_URL` (default: `http://worker-api:8090`)
 - `Optional`: `HEALTHCHECK_PORT` (default: `3000`)
 - `Optional`: `DISCORD_SENDMSG_CHARACTER_LIMIT` (default: `2000`)
 - `Optional`: `CHECK_EMAIL_WAIT` (default: `2`)
@@ -174,7 +181,7 @@ Use `.env.example` as the source of truth for defaults.
 
 ### Discord CRM Audit Logging (Best Effort)
 
-- `Optional`: `AUDIT_API_BASE_URL` (when set with `WEBHOOK_SHARED_SECRET`, CRM commands emit best-effort audit events)
+- `Optional`: `AUDIT_API_BASE_URL` (when set with `API_SHARED_SECRET`, CRM commands emit best-effort audit events)
 - `Optional`: `AUDIT_API_TIMEOUT_SECONDS` (default: `2.0`)
 
 ### Kimai (Legacy/Deprecating)