get_limits
Return plan limits for the current API key (GET /v1/limits).
LABSCO SUMMARY
A validation call that explicitly does not consume quota is the design decision that makes a metered scoring API safe to hand an agent.
The failure mode for metered scoring is an agent iterating on payload shape and burning the month's allowance on rejected requests. Separating validation from scoring, and saying in the tool text that validation is free, removes it — the agent can converge on a correct payload and then spend once. get_workflow_schema serves the same end from the other direction, by supplying the field list rather than leaving it to be inferred. Reporting model versions on list_workflows is worth noting for anyone keeping scores comparable over time: a score without its model version is not comparable to last quarter's.
What it is
A workflow-scoring client with eight tools: the scoring calls themselves, the schema and validation around them, and the plan limits and usage reporting that govern them.
What you get
- score_workflow scores one item; score_batch handles up to 25 sequentially in a single request.
- validate_workflow_payload validates and normalises a payload and states that it does not consume a scoring call — so a malformed request costs nothing to discover.
- get_workflow_schema returns required fields, optional fields and example payloads for a workflow, which is what makes a first attempt correct.
- list_workflows returns the supported workflows with their current model versions, so a score can be attributed to a version rather than to 'the model'.
- get_limits and get_usage report the plan's ceiling and the consumption against it, with usage optionally scoped to a month.
- submit_feedback attaches structured feedback to a prior scoring request rather than to the model in general.
Requirements
SENTINEL_BASE_URL and SENTINEL_TOKEN_BASE_URL — the base URLs are configurable rather than fixed, so this points at the deployment you are entitled to.
Setup effort
One command — uvx sentinel-signal-mcp
USE CASES
01
Getting the payload right before spending anything
Read the workflow's schema with its example payloads, then validate and normalise your version — neither call consumes a scoring credit.
02
Scoring a set of items
Send up to 25 in one batch request rather than looping single calls, once the shape is confirmed.
03
Staying inside the plan
Check the ceiling and the consumption against it, scoped by month when you need the trend rather than the total.
04
Telling them a score was wrong
Attach structured feedback to the specific scoring request, so it is tied to an input rather than filed as a general complaint.
8 tools
Counted once with only startup setup. Other versions, modes or optional credentials may expose a different list.
get_usage
Return usage for the current API key (GET /v1/usage). Optional month format: YYYY-MM.
get_workflow_schema
Fetch required fields, optional fields, and example payloads for a workflow (GET /v1/workflows/{workflow}/schema).
list_workflows
List supported workflows and current model versions (GET /v1/workflows).
score_batch
Score up to 25 workflow items sequentially in one request (POST /v1/score/batch).
score_workflow
Call Sentinel Signal unified workflow scoring (POST /v1/score).
submit_feedback
Submit structured feedback for a prior scoring request (POST /v1/feedback).
validate_workflow_payload
Validate and normalize a workflow payload without consuming a scoring call (POST /v1/workflows/{workflow}/validate).
BEFORE YOU INSTALL
What to know first
Tools you get
8 tools when we connected. When your key's quota runs out, tools return a 402 upgrade_required error instead of disappearing.
Requires
uvx on PATH
CONTEXT COST
Light Standing cost
Rated by the size of the 8 tool definitions that ride along on every turn, not by how many there are
INSTALLATION
Connect your client
Server setup
uvx
Any MCP client — stdio
TERMINAL
uvx sentinel-signal-mcpHosted endpoint
ENDPOINT URL
https://mcp.sentinelsignal.io/mcpAdd this hosted endpoint to your MCP client.
MCP Clients
Claude Desktop
CLAUDE DESKTOP · JSON
{
"mcpServers": {
"sentinel-signal": {
"command": "uvx",
"args": [
"sentinel-signal-mcp"
],
"env": {
"SENTINEL_BASE_URL": "https://api.sentinelsignal.io",
"SENTINEL_TOKEN_BASE_URL": "https://token.sentinelsignal.io",
"SENTINEL_API_KEY": "<YOUR_SENTINEL_API_KEY>",
"SENTINEL_TIMEOUT_SECONDS": "30"
}
}
}
}SENTINEL_API_KEY is optional — delete that line and the server mints a free trial key on first run (no signup) and caches it in ~/.sentinel/credentials.json.
Cursor, Windsurf
JSON
{
"mcpServers": {
"sentinel-signal": {
"command": "uvx",
"args": [
"sentinel-signal-mcp"
],
"env": {
"SENTINEL_BASE_URL": "https://api.sentinelsignal.io",
"SENTINEL_TOKEN_BASE_URL": "https://token.sentinelsignal.io"
}
}
}
}Any MCP client — HTTP
Connect a client that supports remote HTTP MCP to this address:
ANY MCP CLIENT — HTTP · URL
https://mcp.sentinelsignal.io/mcpSend a key with each request in the header Authorization: Bearer <YOUR_SENTINEL_API_KEY>. Over HTTP the key is not minted for you; a free trial key needs no signup: POST https://token.sentinelsignal.io/v1/keys/trial
