Shared memory and context tools for agentic work.
Code Rooms
{
"recorded_at": "2026-03-25T00:23:23.032084+00:00",
"scenario_id": "warm_trail_resume_stale_force_recovery",
"scenario_name": "Warm recovery from stale trail resume gate",
"scenario_tags": [
"warm-graph",
"recovery",
"trail",
"continuity",
"stale"
],
"mode": "m1nd_warm",
"execution_origin": "replay",
"source_ref": "docs/benchmarks/events/warm-trail-resume-stale-force-recovery.json",
"cold_graph_time_ms": null,
"warm_graph_time_ms": null,
"time_to_first_good_answer_ms": 1.0,
"time_to_full_proof_ms": 2.6,
"answer_quality": "high",
"plan_changed": false,
"false_start_count": 0,
"tests_identified_before_edit": 0,
"public_claim_worthy": true,
"workflow_notes": "trail_resume teaches the force=true retry directly, then restores partial continuity with a next-step hint.",
"notes": "",
"events": [
"tool_name": "m1nd.trail_resume",
"query": "resume stale trail without force",
"elapsed_ms": 1.0,
"payload_chars": 220,
"hint": "Trail trail_benchmark is stale: 8 of 12 nodes missing (67%). Use force=true to resume.",
"suggested_next_step": "Retry trail_resume with force=true if partial continuity is still useful.",
"usage_hint": "Use trail_resume to continue a saved investigation. If the trail is stale but still valuable, retry with force=true.",
"next_suggested_tool": "trail_resume",
"next_tool_used": "trail_resume",
"recovery_followed": true,
"notes": "The stale gate now teaches when to accept degraded continuity instead of restarting the investigation from zero.",
"event_index": 1,
"opened_files": [],
"surfaced_files": []
},
"query": "retry trail_resume with force=true",
"elapsed_ms": 1.4,
"payload_chars": 360,
"next_suggested_tool": "view",
"next_suggested_target": "m1nd-mcp/src/layer_handlers.rs",
"next_step_hint": "Open the restored focus next to confirm whether the degraded continuity still answers the carried-forward question.",
"next_tool_used": "view",
"notes": "Forced resume restores partial continuity with explicit next-step guidance.",
"event_index": 2,
"tool_name": "m1nd.view",
"target": "m1nd-mcp/src/layer_handlers.rs",
"elapsed_ms": 0.7,
"payload_chars": 520,
"opened_files": [
"m1nd-mcp/src/layer_handlers.rs"
"event_index": 3,
}
"files_opened": 1,
"repeat_reads": 0,
"search_iterations": 0,
"chars_surfaced": 1100,
"token_proxy": 275,
"guidance_events": 2,
"guidance_followed": 2,
"guidance_followthrough_rate": 1.0,
"reactivated_nodes": 0,
"resume_hints": 0,
"proof_states": [],
"final_proof_state": null,
"proof_state_counts": {},
"progress_events": 0,
"max_progress_pct": 0.0,
"active_phases": [],
"next_phases": [],
"progress_event_types": [],
"progress_delivery_modes": [],
"live_progress_events": 0,
"replay_progress_events": 0,
"snapshot_progress_events": 0,
"progress_guidance_events": 0,
"progress_guidance_followed": 0,
"progress_guidance_followthrough_rate": null,
"recovery_events": 1,
"recovery_followed": 1,
"recovery_followthrough_rate": 1.0,
"repo_path": null,
"question": "Recover from trail_resume rejecting a stale trail with too many missing nodes, then retry with force=true and continue from the degraded but still useful continuity state.",
"expected_strength": "strong"