feat: add monitoring.yaml as source of truth for protocol cards

CLAUDE.md

@@ -37,6 +37,13 @@
 - **Logging**: Use structured logging instead of print statements
 - **Testing**: Write unit tests for utility functions and integration tests for protocols
 
+## Monitoring Metadata
+
+- `monitoring.yaml` at the repo root is the source of truth for the protocol cards on https://curation.yearn.fi/monitoring/.
+- When adding, removing, or changing a monitor, update `monitoring.yaml` and the protocol `README.md`.
+- The alerts API exposes this metadata at `GET /v1/monitoring` (see `api/server.py`).
+- `tests/test_monitoring_config.py` validates that every enabled protocol task in `automation/jobs.yaml` has a `monitoring.yaml` entry and that every task referenced there is scheduled.
+
 ## Best Practices
 - Create custom exception types for domain-specific errors
 - Extract repeated patterns into utility functions. Try to keep utility functions small and focused on a single task and store in utils folder.

CONTRIBUTING.md

@@ -149,9 +149,10 @@ write_last_value_to_file(cache_filename, "MY_KEY", new_value)
 
 1. Create `protocols/protocol-name/main.py` following the pattern above, with a `main()` function and an `if __name__ == "__main__":` block that wraps it via `run_with_alert(main, PROTOCOL)` (see [Script Entrypoint](#script-entrypoint)). Reference ABIs by their repo-root-relative path, e.g. `load_abi("protocols/protocol-name/abi/Foo.json")`
 2. Add a `protocols/protocol-name/README.md` describing what it monitors
-3. No `pyproject.toml` change is needed — packages under `protocols/` are discovered automatically (see `[tool.setuptools.packages.find]`)
-4. Add the corresponding `TELEGRAM_BOT_TOKEN_*` and `TELEGRAM_CHAT_ID_*` entries to `.env.example`
-5. Register the script in `automation/jobs.yaml` under the right profile if it should run on a schedule
+3. Add or update the protocol entry in `monitoring.yaml` so the curation website card stays in sync
+4. No `pyproject.toml` change is needed — packages under `protocols/` are discovered automatically (see `[tool.setuptools.packages.find]`)
+5. Add the corresponding `TELEGRAM_BOT_TOKEN_*` and `TELEGRAM_CHAT_ID_*` entries to `.env.example`
+6. Register the script in `automation/jobs.yaml` under the right profile if it should run on a schedule
 
 ## Tests
 

README.md

@@ -80,7 +80,7 @@ uv run protocols/aave/main.py
 
 In production the scripts run on a schedule via supercronic on a VPS, defined in [`automation/jobs.yaml`](./automation/jobs.yaml). See [`deploy/`](./deploy/) — [`install.sh`](./deploy/install.sh) provisions a host and [`runbook.md`](./deploy/runbook.md) covers operations.
 
-The optional read-only alerts API exposes persisted alert history from SQLite. See [`deploy/alerts-api.md`](./deploy/alerts-api.md) for endpoint examples and pagination.
+The optional read-only alerts API exposes persisted alert history from SQLite and monitoring card metadata. See [`deploy/alerts-api.md`](./deploy/alerts-api.md) for endpoint examples and pagination, or `GET /v1/monitoring` for the protocol card data used by the curation website.
 
 ## Code Style
 

api/server.py

@@ -9,6 +9,7 @@
 
 from automation.config import JobsConfig, load_jobs_config
 from utils.logger import get_logger
+from utils.monitoring_config import load_monitoring_config, monitoring_to_json
 from utils.store import AlertEvent, get_alert, normalize_timestamp, query_alerts
 
 logger = get_logger("api.server")
@@ -160,6 +161,9 @@ def do_GET(self) -> None:
             if parsed.path == "/v1/protocols":
                 write_json(self, 200, protocols_to_json(load_jobs_config()))
                 return
+            if parsed.path == "/v1/monitoring":
+                write_json(self, 200, monitoring_to_json(load_monitoring_config()))
+                return
             if parsed.path == "/v1/alerts":
                 alert_query = parse_alert_query(parsed.query)
                 rows = query_alerts(

deploy/alerts-api.md

@@ -130,6 +130,41 @@ Example response:
 }
 ```
 
+### `GET /v1/monitoring`
+
+Returns the structured protocol card metadata from `monitoring.yaml`, used by
+the curation website to render what each protocol monitors.
+
+```sh
+curl http://127.0.0.1:8923/v1/monitoring
+```
+
+Example response:
+
+```json
+{
+  "version": "1.0",
+  "data": [
+    {
+      "slug": "3jane",
+      "display_name": "3Jane",
+      "description": "USD3/sUSD3 credit market monitoring on Mainnet",
+      "cadence": "Hourly",
+      "monitor_count": 11,
+      "disabled": false,
+      "tasks": ["protocols/3jane/main.py"],
+      "monitors": [
+        {
+          "name": "Price Per Share",
+          "description": "USD3 PPS decrease (any drop vs cached prior, CRITICAL) and sUSD3 PPS decrease (HIGH)"
+        }
+      ]
+    }
+  ],
+  "count": 1
+}
+```
+
 ## Delivery Status
 
 An alert row means a monitor generated an alert. Telegram delivery is tracked