docs: add missing env vars for image search and telemetry

README.md

@@ -79,18 +79,32 @@ All commands return JSON to stdout.
 
 ## Environment variables
 
+Copy `.env.example` to `.env` and fill in the values.
+
+### Vision (required for `detect`)
+
 | Variable | Required | Description |
 |----------|----------|-------------|
-| `VISION_API_KEY` | Yes | API key for the vision model endpoint |
+| `VISION_API_KEY` | Yes | API key for the vision model |
 | `VISION_API_BASE` | No | OpenAI-compatible base URL (default: OpenAI) |
 | `VISION_MODEL` | No | Model name (default: `gpt-4o-mini`) |
 
-```bash
+### Image search (required for `search` / `detect-and-search`)
-# Use a local or custom provider
+
-VISION_API_BASE=https://your-llm-endpoint/v1
+| Variable | Required | Description |
-VISION_MODEL=claude-haiku-4-5-20251001
+|----------|----------|-------------|
-VISION_API_KEY=sk-...
+| `ONEBOUND_UPLOAD_ENDPOINT` | Yes | Endpoint to upload a local image and get a public URL |
-```
+| `ONEBOUND_SEARCH_ENDPOINT` | Yes | Reverse image search endpoint |
+| `ONEBOUND_KEYWORD_SEARCH_ENDPOINT` | No | Keyword search endpoint for re-ranking results |
+
+These proxy through a local `woo-data-scrawler` instance — no Onebound API key needed directly.
+
+### Other
+
+| Variable | Required | Description |
+|----------|----------|-------------|
+| `AUTH_RT_BIN` | No | Override path to the `auth-rt` binary |
+| `TELEMETRY_ENDPOINT` | No | POST skill execution results to a Loki-compatible endpoint |
 
 ## Prerequisites