{
  "batch": "deep-41",
  "setup": "Cloned Hermes source under test/scratch/hermes and ran uv sync --locked with a local CPython 3.14.10; the CLI exposed install, validate, and doctor. tryit.py empty-plan runs were attempted with a 25-second bound but did not emit a report, so tool-call execution could not be observed.",
  "plugins": [
    {
      "pr": 0,
      "name": "done-bell",
      "validate": "passed",
      "inside": "Not determined: the bounded empty-plan tryit.py run emitted no report, so no added tools or calls were observed.",
      "tests": "passed, 9 passed",
      "useful": 3,
      "reliable": 5,
      "polish": 4,
      "creative": 2,
      "beyond": 3,
      "total": 20,
      "rank": 8,
      "best": "A small, privacy-conscious completion notifier with passing tests.",
      "worst": "It improves notification ergonomics but does not materially extend agent capability.",
      "blocked": null
    },
    {
      "pr": 0,
      "name": "homeassistant",
      "validate": "passed",
      "inside": "Not determined: the bounded empty-plan tryit.py run emitted no report, so no added tools or calls were observed.",
      "tests": "not run; no focused suite was run",
      "useful": 5,
      "reliable": 4,
      "polish": 4,
      "creative": 3,
      "beyond": 5,
      "total": 26,
      "rank": 5,
      "best": "A mature platform adapter gives Hermes real smart-home events and control.",
      "worst": "It requires a configured Home Assistant service and access token.",
      "blocked": null
    },
    {
      "pr": 151968,
      "name": "tickover",
      "validate": "failed: capability probe initially raised No module named 'numpy'; manifest, loadability, dependency declaration, security scan, and collision checks passed.",
      "inside": "Not determined: the bounded empty-plan tryit.py run did not emit a report after heavyweight dependency setup, so no tool calls were observed.",
      "tests": "not completed: 9 tests passed before the 25-second bound",
      "useful": 5,
      "reliable": 3,
      "polish": 5,
      "creative": 5,
      "beyond": 5,
      "total": 28,
      "rank": 2,
      "best": "It gives Hermes local, testable acoustic verification and drift monitoring for physical actions.",
      "worst": "Its large runtime/model dependency path prevented a complete in-Hermes run here.",
      "blocked": null
    },
    {
      "pr": 151887,
      "name": "cmi",
      "validate": "passed",
      "inside": "Not determined: the bounded empty-plan tryit.py run emitted no report, so no hooks or calls were observed in conversation.",
      "tests": "failed, 314 passed, 3 skipped, 1 failed (concurrent process-writer test)",
      "useful": 5,
      "reliable": 4,
      "polish": 5,
      "creative": 5,
      "beyond": 5,
      "total": 29,
      "rank": 1,
      "best": "A deterministic context engine changes how Hermes retains and compacts long-running agent work.",
      "worst": "One concurrent persistence test failed in the supplied environment.",
      "blocked": null
    },
    {
      "pr": 153524,
      "name": "shadow-army",
      "validate": "passed",
      "inside": "Not determined: the bounded empty-plan tryit.py run emitted no report, so no hooks or calls were observed in conversation.",
      "tests": "failed, 3 passed, 1 failed (empty-state setup count)",
      "useful": 4,
      "reliable": 3,
      "polish": 5,
      "creative": 4,
      "beyond": 4,
      "total": 24,
      "rank": 7,
      "best": "Independent-model verification before closing code work is a thoughtful anti-reward-hacking guardrail.",
      "worst": "It depends on a separately configured verifier route, and its suite had an early failure.",
      "blocked": null
    },
    {
      "pr": 152214,
      "name": "passkey-gate",
      "validate": "passed after its declared webauthn and qrcode dependencies were installed",
      "inside": "Not determined: the bounded empty-plan tryit.py run emitted no report, so no gate tools or approval calls were observed.",
      "tests": "failed, 1 passed, 1 failed (audit-notarize test)",
      "useful": 4,
      "reliable": 3,
      "polish": 5,
      "creative": 4,
      "beyond": 5,
      "total": 26,
      "rank": 4,
      "best": "It binds risky agent actions to a person\u2019s passkey approval of the exact action.",
      "worst": "It needs phone/HTTPS setup and its audit suite failed early in this environment.",
      "blocked": null
    },
    {
      "pr": 152256,
      "name": "nvidia-switchyard",
      "validate": "passed",
      "inside": "Not determined: the bounded empty-plan tryit.py run emitted no report, so no provider route or calls were observed.",
      "tests": "failed, 9 passed, 2 skipped, 1 failed (Hermes side-task provider parity)",
      "useful": 4,
      "reliable": 3,
      "polish": 5,
      "creative": 4,
      "beyond": 5,
      "total": 26,
      "rank": 6,
      "best": "Local-first verification-gated routing changes Hermes\u2019s model platform rather than adding another prompt workflow.",
      "worst": "It is explicitly beta, requires local and cloud model configuration, and had a parity-test failure.",
      "blocked": null
    },
    {
      "pr": 153809,
      "name": "living-spaces",
      "validate": "passed",
      "inside": "Not determined: the bounded empty-plan tryit.py run emitted no report, so no Space tools or calls were observed.",
      "tests": "none found in the plugin package; not run",
      "useful": 4,
      "reliable": 3,
      "polish": 5,
      "creative": 5,
      "beyond": 5,
      "total": 27,
      "rank": 3,
      "best": "Persistent, sandboxed canvases with records, revisions, Desktop UI, and Hermes upkeep are a distinctive platform extension.",
      "worst": "The package documents dependence on private cron and Desktop interfaces and carries no executable package tests.",
      "blocked": null
    }
  ]
}