Measurement
Information that was previously trapped in JSON files or trace summaries is now surfaced as operator packets.
The operator shell is justified by real packets and workflows, not by generic feature count.
| Use case | Type | Current output |
|---|---|---|
| 1. Benchmark Review | Measurement | Deterministic benchmark review packet from benchmark_1_10_results.json. |
| 2. Run-Trace Triage | Measurement | Replay-target and escalation packet from the trace summaries. |
| 3. Incident Repair Loop | Operator discipline | Structured incident create → repair → resolve → validate workflow. |
| 4. Model Comparison Packet | Measurement | Capture-integrity-aware model packet for the protocol lane. |
| 5. Documentation Status Review | Operator discipline | Deterministic doc drift detector for the OpenClaw/Phoenix doc set. |
| 6. Routing Policy | Decision surface | Explicit four-lane routing policy grounded in the benchmark line. |
| 7. Solver-Backed Demo Expansion | Decision surface / demo | ShowcaseAgent routing demo confirms the rule-surface argument in a second domain. |
Information that was previously trapped in JSON files or trace summaries is now surfaced as operator packets.
The shell now enforces hardening, incident handling, and documentation review rather than relying on memory and ad hoc checks.
Phoenix findings are now queryable by operators without requiring deep internal repo knowledge.
The use-case bundle is paired with the operator-shell architecture paper. The paper argues the pattern; the use cases prove the shell reduced friction without crossing the authority boundary.