[analysis]
Compares two stored GliderTrace sessions and highlights fixed, new, and persisting evidence.
Verification work depends on knowing what was fixed, what is new, and what still persists.
The model can compare two evidence sessions directly instead of manually diffing logs.
Try it locally in the GliderTrace playground
| Name | Type | Required | Description |
|---|---|---|---|
| baselineSessionId | string | Yes | Baseline trace session id, usually before a fix or change. |
| candidateSessionId | string | Yes | Candidate trace session id, usually after a fix or change. |
| kinds | json | No | Optional evidence kinds such as testFailure, exception, log, counter, process, artifact, or warning. |
| minSeverity | string | No | Optional minimum severity: 'info', 'warning', 'error', or 'critical'. |
| limit | number | No | Maximum number of fixed, new, and persisting items to return per group. |
| includeArtifacts | boolean | No | Include artifact references from both sessions. Defaults to true. |
Compare before and after sessions
{
"jsonrpc": "2.0",
"id": 1,
"method": "tools/call",
"params": {
"name": "trace_compare_sessions",
"arguments": {
"baselineSessionId": "trc_before",
"candidateSessionId": "trc_after"
}
}
}Returns fixed, new, and persisting evidence groups with optional artifact references.