Tool selection guide
When structured runtime evidence helps
GliderTrace captures and stores runtime evidence for .NET coding agents. It complements the .NET CLI and does not replace it.
Use it when an agent must query a run after the command ends, relate evidence to artifacts, or compare two sessions.
Choose GliderTrace from the evidence
Commands and tests
Run a workspace-scoped command and retain its evidence.
trace_runtrace_buildtrace_restoretrace_run_tests- Sessions retain exit codes, stdout, stderr, and normalized findings.
- Test runs can retain TRX artifacts and identified failures.
Long-running processes
Give a service or application an explicit evidence boundary.
trace_starttrace_stop- Start and stop one stored session around the process.
- Inspect the finalized session after the process stops.
Existing artifacts
Bring CI or local artifacts into a queryable session.
trace_import_artifactstrace_import_otlptrace_export- Import supported logs, TRX, counters, traces, dumps, or OpenTelemetry data.
- Exports can reference local artifacts instead of placing raw binary data in a prompt.
Counters, traces, and dumps
Collect explicit runtime diagnostics from a local .NET process.
trace_attachtrace_counterstrace_hotspotstrace_capture_dumptrace_capture_gcdump- Sensitive dump capture requires explicit acknowledgement.
- Findings retain their session and artifact identity.
Session queries
Read a bounded part of stored evidence after a run.
trace_get_sessiontrace_query_eventstrace_exceptions- Filter normalized events by kind, text, severity, or cursor.
- Group exception evidence by type, message, and first stack frame.
Before and after
Compare two stored sessions after a code change.
trace_compare_sessionstrace_export- Comparison output groups fixed, new, and persisting evidence.
- Keep the underlying commands and session identifiers with the result.
Use another tool when
- Use the .NET CLI directly when a stored session or structured result adds no value.
- Use Glider for compiler-backed C# symbols, relationships, diagnostics, and refactor previews.
- Use Scout for repository retrieval across files and languages.
- This page makes no claim about fewer calls, lower token use, or shorter task time.
About comparisons
GliderTrace has concrete storage, query, comparison, and export capabilities. Those facts do not prove a workflow advantage.
Any future agent-workflow comparison must publish its commands, artifacts, prompts, models, scoring rules, and raw outcomes.