Groundline
Server Configuration
Describes the environment variables required to run the server.
| Name | Required | Description | Default |
|---|---|---|---|
No arguments | |||
Instructions
Guidance the server publishes about itself, which clients place ahead of the tool catalog so the model reads it before choosing anything.
This server publishes no instructions, or was last inspected before Glama recorded them.
Capabilities
Features and capabilities supported by this server
Protocol revision2025-11-25
| Capability | Details |
|---|---|
| tools | {
"listChanged": false
} |
| prompts | {
"listChanged": false
} |
| resources | {
"subscribe": false,
"listChanged": false
} |
Tools
Functions exposed to the LLM to take actions
| Name | Description |
|---|---|
| open_runA | Load a test run (CSV or TDMS). reference.csv / limits.json next to the file are picked up automatically. Returns a run_id and an overview of the channels. |
| list_analysis_toolsA | List the deterministic analysis tools and their parameter schemas. |
| run_analysisB | Run one analysis tool. The call is stored in the evidence ledger; cite the returned evidence_id in findings. |
| verifyA | Check findings (title, statement, category, severity, channel, t_start, t_end, evidence[]) against the ledger. Every number in a statement must be present in the cited evidence, with a unit and role that fit the evidence field it came from (e.g. a number after "peak" must come from a peak field). A finding in an anomaly category must cite the tool result that reports that anomaly; passed checks are 'observation'. Tag a number with its source field to have it checked against exactly that field: "742.3 K [E4.violations[0].peak_value]", "3 spikes [len(E3.issues[0].spikes)]". |
| write_html_reportC | Verify the findings and write the HTML report (plus report.json for |
Prompts
Interactive templates invoked by user choice
| Name | Description |
|---|---|
No prompts | |
Resources
Contextual data attached and managed by the client
| Name | Description |
|---|---|
No resources | |
TDQS
Scored across 5 tools
Each tool maps to a distinct step: load a run, discover analyses, execute one, verify findings, and emit a report. The only soft overlap is that write_html_report also performs verification, which could cause an agent to skip the standalone verify tool, but the boundaries remain clear.
Most tools follow a verb_noun or verb_object pattern (open_run, list_analysis_tools, run_analysis, write_html_report). The lone bare verb 'verify' is a minor deviation but still readable and unambiguous.
Five tools is well-scoped for a run-analysis pipeline, with each tool earning its place. It is on the lean side but nothing feels redundant.
The surface covers the core lifecycle: open a run, discover analyses, execute them, verify evidence, and produce a report. Minor gaps exist (no way to list previously opened runs or explicitly close a run), but agents can work around these.