{
  "schemaVersion": 1,
  "evidenceClass": "historical_first_party_observation",
  "asOf": "2026-07-11",
  "sourceCommit": "cf4620ea7c931d30d07a965d85af755b98da5dd4",
  "limitations": [
    "One task and one client pair; no general client-parity or current-release claim.",
    "Baseline and OrgX arms used different writer model IDs; this is not a controlled causal comparison.",
    "The outage failure is retained; records existing did not make them retrievable.",
    "No new paid model run was performed for this package."
  ],
  "runs": [
    {
      "benchmark": "agent-amnesia-test",
      "run_date": "2026-07-11",
      "seed_task": "api-rate-limiting",
      "client_a": "claude-code",
      "client_b": "opencode",
      "model_a": "claude-opus-4-1-20250805",
      "model_b": "opencode/big-pickle",
      "with_orgx": false,
      "scores": {
        "decision": false,
        "provenance": false,
        "owner": false,
        "artifact": false,
        "next_action": false
      },
      "score": "0/5",
      "sourceUrl": "https://github.com/useorgx/orgx-mcp/blob/cf4620ea7c931d30d07a965d85af755b98da5dd4/docs/benchmarks/agent-amnesia-test/runs/2026-07-11-claude-code-opencode/run-1-baseline.json"
    },
    {
      "benchmark": "agent-amnesia-test",
      "run_date": "2026-07-11",
      "seed_task": "api-rate-limiting",
      "client_a": "claude-code",
      "client_b": "opencode",
      "model_a": "claude-fable-5",
      "model_b": "opencode/big-pickle",
      "with_orgx": true,
      "scores": {
        "decision": false,
        "provenance": false,
        "owner": false,
        "artifact": false,
        "next_action": false
      },
      "score": "0/5",
      "sourceUrl": "https://github.com/useorgx/orgx-mcp/blob/cf4620ea7c931d30d07a965d85af755b98da5dd4/docs/benchmarks/agent-amnesia-test/runs/2026-07-11-claude-code-opencode/run-2-with-orgx-outage.json"
    },
    {
      "benchmark": "agent-amnesia-test",
      "run_date": "2026-07-11",
      "seed_task": "api-rate-limiting",
      "client_a": "claude-code",
      "client_b": "opencode",
      "model_a": "claude-fable-5",
      "model_b": "opencode/big-pickle",
      "with_orgx": true,
      "scores": {
        "decision": true,
        "provenance": true,
        "owner": true,
        "artifact": true,
        "next_action": true
      },
      "score": "5/5",
      "sourceUrl": "https://github.com/useorgx/orgx-mcp/blob/cf4620ea7c931d30d07a965d85af755b98da5dd4/docs/benchmarks/agent-amnesia-test/runs/2026-07-11-claude-code-opencode/run-3-with-orgx.json"
    }
  ]
}