[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"skill:01M73DTF5CD5JW5R5X8F83H72R":3,"skill-same:01M73DTF5CD5JW5R5X8F83H72R":47},{"skillNo":4,"name":5,"nameZh":6,"description":6,"descriptionZh":6,"repositoryFullName":7,"ownerLogin":8,"ownerAvatarUrl":9,"ownerAccountNo":10,"ownerDisplayName":11,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":15,"categories":16,"keywords":17,"publishedAt":18,"linkOnly":19,"markdownContent":20,"markdownBody":21,"frontMatter":22,"refName":23,"commitSha":24,"skillDirectory":25,"forksCount":26,"repositoryHtmlUrl":27,"repositoryHomepageUrl":28,"ownerHtmlUrl":29,"ownerBlogUrl":30,"ownerDescription":31,"updatedAt":32,"install":33,"resources":46},"01M73DTF5CD5JW5R5X8F83H72R","mlperf-inference-benchmarks",null,"Horonomos-laps\u002FYana-AI","Horonomos-laps","https:\u002F\u002Favatars.githubusercontent.com\u002Fu\u002F355299122?v=4","01M76B3QPXE88ET2SQPDPVATZT"," Horonomos-laps","01M72ZQRJPCYPSSWT4D00VJR2A",13,"Apache-2.0","https:\u002F\u002Fgithub.com\u002FHoronomos-laps\u002FYana-AI\u002Fblob\u002Fmain\u002Fcore\u002Fskills\u002Fmlperf-inference-benchmarks\u002FSKILL.md",[],[],"2026-11-11 15:11:14",false,"---\nname: mlperf-inference-benchmarks\ndescription: MLPerf inference benchmarking patterns. Loadgen scenarios (SingleStream\u002FMultiStream\u002FServer\u002FOffline), SUT\u002FQSL interfaces, latency p99 targets, and performance metric interpretation. Sources: mlcommons\u002Finference (Apache-2.0).\norigin: yana-ai — synthesized from mlcommons\u002Finference (Apache-2.0)\nlicense: Apache-2.0\nversion: 1.0.0\ncompatibility: yana-ai >= 1.3.51\n---\n\n# \u002Fmlperf-inference-benchmarks\n\n## When to Use\n\n- Measure inference system performance with industry-standard reproducible metrics\n- Compare GPU\u002FCPU hardware choices for a given model\u002Ftask\n- Validate SLA targets: p99 latency \u003C 100ms at 1000 QPS\n- Pre-deployment stress test: flood the system to find breaking point\n\n## Do NOT use for\n\n- Training benchmarks (use MLPerf Training)\n- Quick sanity checks (use [[llm-eval-framework]])\n\n---\n\n## MLPerf scenarios\n\n```\nSingleStream: one query at a time, measure latency distribution\n  Target: p90 latency (e.g., \u003C 10ms for image classification)\n  Use: edge devices, latency-sensitive APIs\n\nMultiStream: N queries in parallel, measure all-complete latency\n  Target: p99 latency across all N queries\n  Use: camera pipelines, fixed batch size hardware\n\nServer: Poisson-distributed query arrival, measure sustained throughput\n  Target: max QPS while keeping p99 \u003C SLA threshold\n  Use: production cloud inference\n\nOffline: all queries available upfront, measure total throughput\n  Target: max queries per second (no latency constraint)\n  Use: batch processing, overnight jobs\n```\n\n---\n\n## Custom Loadgen harness (Python)\n\n```python\nimport mlperf_loadgen as lg\n\nclass YamtamSUT:\n  \"\"\"System Under Test — wraps your inference engine\"\"\"\n\n  def issue_queries(self, query_samples):\n    for qs in query_samples:\n      # Run inference\n      result = run_inference(qs.index)\n      # Report response\n      response = lg.QuerySampleResponse(\n        id     = qs.id,\n        data   = result.ctypes.data,\n        size   = result.nbytes,\n      )\n      lg.QuerySamplesComplete([response])\n\n  def flush_queries(self): pass\n\n\nclass YamtamQSL:\n  \"\"\"Query Sample Library — dataset loader\"\"\"\n\n  def __init__(self, dataset, total_count, perf_count):\n    self.qsl = lg.ConstructQSL(\n      total_count, perf_count,\n      self.load_query_samples, self.unload_query_samples\n    )\n    self.dataset = dataset\n\n  def load_query_samples(self, sample_list):\n    for s in sample_list: self.dataset.load(s)\n\n  def unload_query_samples(self, sample_list):\n    for s in sample_list: self.dataset.unload(s)\n\n\n# Run benchmark\nsettings = lg.TestSettings()\nsettings.scenario    = lg.TestScenario.Server\nsettings.mode        = lg.TestMode.PerformanceOnly\nsettings.server_target_qps = 500\n\nsut = YamtamSUT()\nqsl = YamtamQSL(dataset, total_count=5000, perf_count=1024)\n\nlg.StartTest(sut, qsl.qsl, settings)\nlg.DestroyQSL(qsl.qsl)\n```\n\n---\n\n## Key metrics to track\n\n```bash\n# After loadgen run, parse mlperf_log_summary.txt\ngrep -E \"(Samples per second|Min latency|Max latency|Mean latency|p99.00 latency)\" \\\n  mlperf_log_summary.txt\n\n# Target thresholds for yamtam tool calls:\n# Mean latency:   \u003C 200ms  (acceptable for interactive agent)\n# p99 latency:    \u003C 500ms  (hard SLA)\n# Throughput:     > 50 QPS (for 50 concurrent agents)\n```\n\n---\n\n## Quick server scenario stress test (without full MLPerf)\n\n```bash\n# Use wrk2 for HTTP inference endpoint load testing\nwrk2 -t 4 -c 100 -d 60s -R 500 \\\n  --script post.lua \\\n  http:\u002F\u002Finference-server:8000\u002Finfer\n\n# post.lua\nwrk.method  = \"POST\"\nwrk.headers = {[\"Content-Type\"] = \"application\u002Fjson\"}\nwrk.body    = '{\"input\": \"benchmark query\"}'\n```\n\n---\n\n## Anti-Fake-Pass Checklist\n\n```\n❌ Using accuracy mode results for performance reporting → different sample count, not valid perf number\n❌ Warmup not included → first queries hit cold cache; p99 inflated\n❌ p50 latency meets SLA but p99 does not → tail latency the real constraint; always report p99\n❌ QPS measured on a single sample → not representative; use Server scenario with realistic load\n❌ Not isolating benchmark host → background processes inflate latency; dedicate hardware\n❌ Comparing results across different MLPerf versions → LoadGen version affects timing methodology\n```\n","# \u002Fmlperf-inference-benchmarks\n\n## When to Use\n\n- Measure inference system performance with industry-standard reproducible metrics\n- Compare GPU\u002FCPU hardware choices for a given model\u002Ftask\n- Validate SLA targets: p99 latency \u003C 100ms at 1000 QPS\n- Pre-deployment stress test: flood the system to find breaking point\n\n## Do NOT use for\n\n- Training benchmarks (use MLPerf Training)\n- Quick sanity checks (use [[llm-eval-framework]])\n\n---\n\n## MLPerf scenarios\n\n```\nSingleStream: one query at a time, measure latency distribution\n  Target: p90 latency (e.g., \u003C 10ms for image classification)\n  Use: edge devices, latency-sensitive APIs\n\nMultiStream: N queries in parallel, measure all-complete latency\n  Target: p99 latency across all N queries\n  Use: camera pipelines, fixed batch size hardware\n\nServer: Poisson-distributed query arrival, measure sustained throughput\n  Target: max QPS while keeping p99 \u003C SLA threshold\n  Use: production cloud inference\n\nOffline: all queries available upfront, measure total throughput\n  Target: max queries per second (no latency constraint)\n  Use: batch processing, overnight jobs\n```\n\n---\n\n## Custom Loadgen harness (Python)\n\n```python\nimport mlperf_loadgen as lg\n\nclass YamtamSUT:\n  \"\"\"System Under Test — wraps your inference engine\"\"\"\n\n  def issue_queries(self, query_samples):\n    for qs in query_samples:\n      # Run inference\n      result = run_inference(qs.index)\n      # Report response\n      response = lg.QuerySampleResponse(\n        id     = qs.id,\n        data   = result.ctypes.data,\n        size   = result.nbytes,\n      )\n      lg.QuerySamplesComplete([response])\n\n  def flush_queries(self): pass\n\n\nclass YamtamQSL:\n  \"\"\"Query Sample Library — dataset loader\"\"\"\n\n  def __init__(self, dataset, total_count, perf_count):\n    self.qsl = lg.ConstructQSL(\n      total_count, perf_count,\n      self.load_query_samples, self.unload_query_samples\n    )\n    self.dataset = dataset\n\n  def load_query_samples(self, sample_list):\n    for s in sample_list: self.dataset.load(s)\n\n  def unload_query_samples(self, sample_list):\n    for s in sample_list: self.dataset.unload(s)\n\n\n# Run benchmark\nsettings = lg.TestSettings()\nsettings.scenario    = lg.TestScenario.Server\nsettings.mode        = lg.TestMode.PerformanceOnly\nsettings.server_target_qps = 500\n\nsut = YamtamSUT()\nqsl = YamtamQSL(dataset, total_count=5000, perf_count=1024)\n\nlg.StartTest(sut, qsl.qsl, settings)\nlg.DestroyQSL(qsl.qsl)\n```\n\n---\n\n## Key metrics to track\n\n```bash\n# After loadgen run, parse mlperf_log_summary.txt\ngrep -E \"(Samples per second|Min latency|Max latency|Mean latency|p99.00 latency)\" \\\n  mlperf_log_summary.txt\n\n# Target thresholds for yamtam tool calls:\n# Mean latency:   \u003C 200ms  (acceptable for interactive agent)\n# p99 latency:    \u003C 500ms  (hard SLA)\n# Throughput:     > 50 QPS (for 50 concurrent agents)\n```\n\n---\n\n## Quick server scenario stress test (without full MLPerf)\n\n```bash\n# Use wrk2 for HTTP inference endpoint load testing\nwrk2 -t 4 -c 100 -d 60s -R 500 \\\n  --script post.lua \\\n  http:\u002F\u002Finference-server:8000\u002Finfer\n\n# post.lua\nwrk.method  = \"POST\"\nwrk.headers = {[\"Content-Type\"] = \"application\u002Fjson\"}\nwrk.body    = '{\"input\": \"benchmark query\"}'\n```\n\n---\n\n## Anti-Fake-Pass Checklist\n\n```\n❌ Using accuracy mode results for performance reporting → different sample count, not valid perf number\n❌ Warmup not included → first queries hit cold cache; p99 inflated\n❌ p50 latency meets SLA but p99 does not → tail latency the real constraint; always report p99\n❌ QPS measured on a single sample → not representative; use Server scenario with realistic load\n❌ Not isolating benchmark host → background processes inflate latency; dedicate hardware\n❌ Comparing results across different MLPerf versions → LoadGen version affects timing methodology\n```\n",{},"main","481dcd7dd2e288b9d627a2bd017db70e64e6c6fa","core\u002Fskills\u002Fmlperf-inference-benchmarks",2,"https:\u002F\u002Fgithub.com\u002FHoronomos-laps\u002FYana-AI","https:\u002F\u002Fyana.vutam.link\u002F","https:\u002F\u002Fgithub.com\u002FHoronomos-laps","https:\u002F\u002Fyana.vutam.link","An open-source governed runtime & execution environment for multi-agent AI systems.","2026-11-14 13:24:09",{"source":7,"skillName":5,"commands":34,"prompt":44,"zipAvailable":45},[35,38,41],{"tool":36,"command":37},"npx","npx skills add Horonomos-laps\u002FYana-AI --skill mlperf-inference-benchmarks",{"tool":39,"command":40},"bunx","bunx skills add Horonomos-laps\u002FYana-AI --skill mlperf-inference-benchmarks",{"tool":42,"command":43},"pnpm","pnpm dlx skills add Horonomos-laps\u002FYana-AI --skill mlperf-inference-benchmarks","安装技能，技能地址：https:\u002F\u002Fgithub.com\u002FHoronomos-laps\u002FYana-AI\u002Fblob\u002Fmain\u002Fcore\u002Fskills\u002Fmlperf-inference-benchmarks\u002FSKILL.md",true,[],[48,56,63,71,79,87],{"skillNo":49,"name":50,"nameZh":6,"description":51,"descriptionZh":6,"repositoryFullName":7,"ownerLogin":8,"ownerAvatarUrl":9,"ownerAccountNo":10,"ownerDisplayName":11,"repositoryNo":12,"starsCount":13,"licenseSpdxId":52,"licenseFileUrl":6,"sourceHtmlUrl":53,"categories":54,"keywords":55,"publishedAt":18},"01M735GKET5YVYKAY62W2P5YXB","openai--sentry--sentry","Use when the user asks to inspect Sentry issues or events, summarize recent production errors, or pull basic Sentry health data via the Sentry API; perform read-only queries with the bundled script and require `SENTRY_AUTH_TOKEN`.","MIT","https:\u002F\u002Fgithub.com\u002FHoronomos-laps\u002FYana-AI\u002Fblob\u002Fmain\u002F.claude\u002Fskills\u002Fopenai--sentry--sentry\u002FSKILL.md",[],[],{"skillNo":57,"name":58,"nameZh":6,"description":59,"descriptionZh":6,"repositoryFullName":7,"ownerLogin":8,"ownerAvatarUrl":9,"ownerAccountNo":10,"ownerDisplayName":11,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":60,"categories":61,"keywords":62,"publishedAt":18},"01M73DCAP83T17HY7SMGK7MGFE","ito-basket-compare","Compare Itô prediction-market baskets against a user's knowledge base, portfolio notes, financial context, watchlist, or research thesis. Use for read-only basket comparison and gap analysis without investment advice or live trading.","https:\u002F\u002Fgithub.com\u002FHoronomos-laps\u002FYana-AI\u002Fblob\u002Fmain\u002Fcore\u002Fskills\u002Fito-basket-compare\u002FSKILL.md",[],[],{"skillNo":64,"name":65,"nameZh":6,"description":66,"descriptionZh":6,"repositoryFullName":7,"ownerLogin":8,"ownerAvatarUrl":9,"ownerAccountNo":10,"ownerDisplayName":11,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":67,"sourceHtmlUrl":68,"categories":69,"keywords":70,"publishedAt":18},"01M734CYEMZFDR3KH6SW0JP8TD","implementing-network-traffic-baselining","Build network traffic baselines from NetFlow\u002FIPFIX data using Python pandas for statistical analysis, z-score anomaly detection, and hourly\u002Fdaily traffic pattern profiling","https:\u002F\u002Fgithub.com\u002Fyanacuti1121\u002FYana-AI\u002Fblob\u002Fmain\u002F.claude\u002Fskills\u002Fimplementing-network-traffic-baselining\u002FLICENSE","https:\u002F\u002Fgithub.com\u002FHoronomos-laps\u002FYana-AI\u002Fblob\u002Fmain\u002F.claude\u002Fskills\u002Fimplementing-network-traffic-baselining\u002FSKILL.md",[],[],{"skillNo":72,"name":73,"nameZh":6,"description":74,"descriptionZh":6,"repositoryFullName":7,"ownerLogin":8,"ownerAvatarUrl":9,"ownerAccountNo":10,"ownerDisplayName":11,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":75,"sourceHtmlUrl":76,"categories":77,"keywords":78,"publishedAt":18},"01M73ASZ6NJRJM44Z2P84NGQ3Q","detecting-evasion-techniques-in-endpoint-logs","Detects defense evasion techniques used by adversaries in endpoint logs including log tampering, timestomping, process injection, and security tool disabling. Use when investigating suspicious endpoint behavior, building detection rules for evasion tactics, or conducting threat hunting for stealthy adversary activity. Activates for requests involving evasion detection, defense evasion analysis, log tampering detection, or MITRE ATT&CK TA0005.","https:\u002F\u002Fgithub.com\u002Fyanacuti1121\u002FYana-AI\u002Fblob\u002Fmain\u002Fcore\u002Fskills\u002Fdetecting-evasion-techniques-in-endpoint-logs\u002FLICENSE","https:\u002F\u002Fgithub.com\u002FHoronomos-laps\u002FYana-AI\u002Fblob\u002Fmain\u002Fcore\u002Fskills\u002Fdetecting-evasion-techniques-in-endpoint-logs\u002FSKILL.md",[],[],{"skillNo":80,"name":81,"nameZh":6,"description":82,"descriptionZh":6,"repositoryFullName":7,"ownerLogin":8,"ownerAvatarUrl":9,"ownerAccountNo":10,"ownerDisplayName":11,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":83,"sourceHtmlUrl":84,"categories":85,"keywords":86,"publishedAt":18},"01M734AFNXD06Z78QRK0VKGT7Q","implementing-microsegmentation-with-guardicore","Implementing microsegmentation using Akamai Guardicore Segmentation to map application dependencies, create granular network policies, visualize east-west traffic flows, and enforce least-privilege communication between workloads across data centers and cloud.","https:\u002F\u002Fgithub.com\u002Fyanacuti1121\u002FYana-AI\u002Fblob\u002Fmain\u002F.claude\u002Fskills\u002Fimplementing-microsegmentation-with-guardicore\u002FLICENSE","https:\u002F\u002Fgithub.com\u002FHoronomos-laps\u002FYana-AI\u002Fblob\u002Fmain\u002F.claude\u002Fskills\u002Fimplementing-microsegmentation-with-guardicore\u002FSKILL.md",[],[],{"skillNo":88,"name":89,"nameZh":6,"description":6,"descriptionZh":6,"repositoryFullName":7,"ownerLogin":8,"ownerAvatarUrl":9,"ownerAccountNo":10,"ownerDisplayName":11,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":90,"categories":91,"keywords":92,"publishedAt":18},"01M736TK27QMCXK0DVCH25SJN5","ramdisk-tmpfs-patterns","https:\u002F\u002Fgithub.com\u002FHoronomos-laps\u002FYana-AI\u002Fblob\u002Fmain\u002F.claude\u002Fskills\u002Framdisk-tmpfs-patterns\u002FSKILL.md",[],[]]