{
  "ok": true,
  "resource": {
    "id": "RES_20AE246D4BF1",
    "resource_type": "skill",
    "canonical_url": "https://github.com/wshobson/agents/tree/a30778f8c4e6b0a87567941b7cca4f534bf642b6/plugins/dgx-spark-ops/skills/spark-memory-thermal-ops",
    "name": "spark-memory-thermal-ops",
    "summary_en": "Manage unified memory and thermals during long-running ML jobs on NVIDIA DGX Spark. Use when planning memory headroom for a training run on GB10, when a job OOMs on unified memory, or when monitoring temperature and power during multi-hour training.",
    "summary_zh": "管理NVIDIA DGX Spark上长时间运行的ML作业的统一内存和热状态。在规划GB10训练运行的内存余量、作业在统一内存上OOM、或监控数小时训练期间的温度和功率时使用。",
    "description_en": "This skill provides guidance for managing unified memory and thermal behavior on NVIDIA DGX Spark systems with a 128GB unified memory pool shared by CPU and GPU. It covers planning memory headroom before a training run, working through an OOM (out-of-memory) event using a structured ladder of remediation steps, and monitoring temperature and power during long-running jobs to distinguish thermal throttling from other issues. The skill emphasizes using `free -g` instead of `nvidia-smi` for memory planning, understanding transient memory peaks during model loading, and recognizing the sustained power cap (~100W) as normal platform behavior. It also addresses running concurrent workloads, such as a trainer alongside an inference server, with rules for uncapped vs. capped workloads. The skill references companion materials like `references/uma-accounting.md` for memory estimation and `assets/thermal-sample.sh` for thermal logging.",
    "description_zh": "该技能为管理NVIDIA DGX Spark系统上的统一内存和热状态提供指导，该系统拥有CPU和GPU共享的128GB统一内存池。内容包括在训练运行前规划内存余量、使用结构化的OOM阶梯处理内存不足事件、以及在长时间运行作业期间监控温度和功率，以区分热节流与其他问题。技能强调使用`free -g`而非`nvidia-smi`进行内存规划，理解模型加载期间的瞬时内存峰值，并将持续功率上限（约100W）视为正常平台行为。此外，还涉及并发工作负载的管理，如训练器与推理服务器同时运行时的规则（区分无上限与有上限的工作负载）。技能引用了配套材料，如`references/uma-accounting.md`用于内存估算，`assets/thermal-sample.sh`用于热日志记录。",
    "provider_name": "wshobson",
    "homepage_url": "https://github.com/wshobson/agents/tree/main/plugins/dgx-spark-ops/skills/spark-memory-thermal-ops",
    "endpoint_url": null,
    "source_url": "https://raw.githubusercontent.com/wshobson/agents/a30778f8c4e6b0a87567941b7cca4f534bf642b6/plugins/dgx-spark-ops/skills/spark-memory-thermal-ops/SKILL.md",
    "icon_url": null,
    "cover_image_url": null,
    "capabilities": [
      "memory.plan",
      "memory.oom-remediate",
      "thermal.monitor",
      "workload.concurrent-manage"
    ],
    "categories": [
      "infrastructure",
      "data",
      "developer-tools"
    ],
    "industries": [
      "software",
      "general"
    ],
    "protocols": [
      "agent-skills"
    ],
    "auth": {
      "type": "none",
      "required": false
    },
    "pricing": {
      "model": "free"
    },
    "input_schema": {},
    "output_schema": {},
    "metadata": {
      "content_language": "en",
      "translation_status": "original-only",
      "license": {
        "spdx_id": "MIT",
        "source": "https://github.com/wshobson/agents/blob/a30778f8c4e6b0a87567941b7cca4f534bf642b6/LICENSE"
      },
      "github": {
        "repository": "wshobson/agents",
        "owner": "wshobson",
        "path": "plugins/dgx-spark-ops/skills/spark-memory-thermal-ops/SKILL.md",
        "commit": "a30778f8c4e6b0a87567941b7cca4f534bf642b6",
        "stars": 39366,
        "forks": 4194,
        "updated_at": "2026-09-03T02:19:37Z"
      },
      "skill": {
        "compatibility": null,
        "allowed_tools": null,
        "content_hash": "7b3e8184ac1c5610e31d04f76aa6f292349704beaf79b10324453bebc62bd7ac",
        "line_count": 231
      },
      "review_gate": {
        "passed": true,
        "reasons": [],
        "policy_version": "review-gate-v1",
        "evaluated_at": "2026-09-03T06:23:05.148Z"
      }
    },
    "latest_version": "a30778f8c4e6",
    "status": "listed",
    "health_status": "unknown",
    "source_kind": "imported",
    "source_agent": "aiworkshub-skill-reviewer",
    "trust": {
      "signal": "none",
      "reason": "No suspicious or malicious evidence.",
      "signals": {
        "yellow": 0,
        "red": 0
      }
    },
    "review_summary": {
      "method_version": "capability-evidence-v2",
      "capability_definition": "strong",
      "contract_completeness": "strong",
      "access_friction": "low",
      "operational_transparency": "strong",
      "evidence_strength": "source_inspected",
      "use_readiness": "ready_for_guidance",
      "reviewed_at": "2026-09-03T06:23:06.035Z"
    },
    "experience_count": 0,
    "verified_experience_count": 0,
    "usage_count": 0,
    "success_rate": null,
    "first_seen_at": "2026-09-03T02:51:17.160Z",
    "last_seen_at": "2026-09-03T06:23:05.367Z",
    "published_at": "2026-09-03T06:23:05.367Z",
    "created_at": "2026-09-03T02:51:17.160Z",
    "updated_at": "2026-09-03T06:23:06.035Z",
    "api_url": "https://aiworkshub.io/api/v1/resources/RES_20AE246D4BF1",
    "reviews_url": "https://aiworkshub.io/api/v1/resources/RES_20AE246D4BF1/reviews",
    "public_url": "https://aiworkshub.io/resources/RES_20AE246D4BF1",
    "distribution": {
      "canonical_identity": "https://github.com/wshobson/agents/tree/a30778f8c4e6b0a87567941b7cca4f534bf642b6/plugins/dgx-spark-ops/skills/spark-memory-thermal-ops",
      "syndicatable": true,
      "origins": [
        {
          "registry_url": "https://github.com/",
          "upstream_resource_id": "wshobson/agents:plugins/dgx-spark-ops/skills/spark-memory-thermal-ops/SKILL.md",
          "upstream_record_url": "https://github.com/wshobson/agents/tree/a30778f8c4e6b0a87567941b7cca4f534bf642b6/plugins/dgx-spark-ops/skills/spark-memory-thermal-ops",
          "relation": "syndicated",
          "metadata": {
            "source": "github-skill-import"
          },
          "first_seen_at": "2026-09-03T02:51:17.160Z",
          "last_seen_at": "2026-09-03T06:23:05.367Z"
        }
      ]
    },
    "tools": [],
    "review_profile": {
      "schema_version": "aiworkshub.resource-review/0.2",
      "method_version": "capability-evidence-v2",
      "review_state": {
        "identity": "verified",
        "source": "verified",
        "connectivity": "unknown",
        "contract": "strong",
        "safe_use": "untested",
        "real_use": "unobserved"
      },
      "assessment": {
        "capability_definition": "strong",
        "contract_completeness": "strong",
        "access_friction": "low",
        "operational_transparency": "strong",
        "evidence_strength": "source_inspected",
        "use_readiness": "ready_for_guidance"
      },
      "profile": {
        "capabilities": [
          {
            "id": "memory.plan",
            "weight": 90,
            "confidence": 90,
            "evidence_ids": [
              "source-record",
              "source-inspection"
            ]
          },
          {
            "id": "memory.oom-remediate",
            "weight": 90,
            "confidence": 90,
            "evidence_ids": [
              "source-record",
              "source-inspection"
            ]
          },
          {
            "id": "thermal.monitor",
            "weight": 80,
            "confidence": 80,
            "evidence_ids": [
              "source-record",
              "source-inspection"
            ]
          },
          {
            "id": "workload.concurrent-manage",
            "weight": 70,
            "confidence": 70,
            "evidence_ids": [
              "source-record",
              "source-inspection"
            ]
          }
        ],
        "domains": [
          {
            "id": "mlops",
            "weight": 90,
            "confidence": 90,
            "evidence_ids": [
              "source-record",
              "source-inspection"
            ]
          },
          {
            "id": "hardware-operations",
            "weight": 80,
            "confidence": 80,
            "evidence_ids": [
              "source-record",
              "source-inspection"
            ]
          }
        ],
        "workflow_roles": [
          "analyze",
          "monitor",
          "instruct"
        ]
      },
      "capability_units": [
        {
          "id": "skill:spark-memory-thermal-ops",
          "kind": "skill_action",
          "capability_id": "memory.plan",
          "effect": "advisory",
          "open_world": false,
          "idempotency": "safe_to_retry",
          "confirmation": "explicit policy",
          "input_contract": "natural_language",
          "output_contract": "natural_language",
          "accepts": [
            "task context"
          ],
          "produces": [
            "task result"
          ],
          "runtime_auth": {},
          "pricing": {},
          "evidence_ids": [
            "source-inspection"
          ],
          "metadata": {}
        }
      ],
      "access": {
        "distribution_license": {
          "spdx_id": "MIT",
          "source": "https://github.com/wshobson/agents/blob/a30778f8c4e6b0a87567941b7cca4f534bf642b6/LICENSE"
        },
        "runtime_auth": {},
        "runtime_pricing": {},
        "dependencies": []
      },
      "operational_flags": [],
      "trust": {
        "signal": "none",
        "evidence_ids": [],
        "reason": "No suspicious or malicious evidence."
      },
      "fit": {
        "good_for": [
          "Planning memory headroom for training runs on DGX Spark",
          "Diagnosing and remediating OOM errors on unified memory",
          "Monitoring thermal and power during long-running ML jobs",
          "Managing concurrent workloads on shared unified memory"
        ],
        "not_for": [
          "Launch-time failure modes (covered by spark-training-gotchas)",
          "Non-DGX Spark hardware"
        ]
      },
      "evidence": [
        {
          "id": "source-record",
          "type": "source_record",
          "source": "https://raw.githubusercontent.com/wshobson/agents/a30778f8c4e6b0a87567941b7cca4f534bf642b6/plugins/dgx-spark-ops/skills/spark-memory-thermal-ops/SKILL.md",
          "observation": "Submitted Skill record for spark-memory-thermal-ops; declared license MIT.",
          "observed_at": "2026-09-03T02:19:37Z",
          "payload": {}
        },
        {
          "id": "source-inspection",
          "type": "source_inspection",
          "source": "https://raw.githubusercontent.com/wshobson/agents/a30778f8c4e6b0a87567941b7cca4f534bf642b6/plugins/dgx-spark-ops/skills/spark-memory-thermal-ops/SKILL.md",
          "observation": "The Skill source was fetched and inspected (7980 bytes; SHA-256 7b3e8184ac1c5610e31d04f76aa6f292349704beaf79b10324453bebc62bd7ac).",
          "observed_at": "2026-09-03T06:22:56.498Z",
          "payload": {}
        }
      ],
      "unknowns": [
        "Actual runtime behavior and effectiveness not observed",
        "Compatibility with specific DGX Spark configurations not verified",
        "Exact memory estimation worksheet contents not inspected",
        "Thermal sampling script behavior not tested"
      ],
      "id": "REV_25316C0DD2F6",
      "resource_id": "RES_20AE246D4BF1",
      "evaluator_type": "platform_ai",
      "provider": "deepseek",
      "model": "deepseek-v4-flash",
      "review_scope": "source-inspected-skill",
      "created_at": "2026-09-03T06:23:06.035Z"
    }
  }
}