{"name":"io.github.Yarmoluk/ckg-nvidia-ai","slug":"yarmoluk-ckg-nvidia-ai","title":"NVIDIA AI CKG","description":"NVIDIA AI knowledge graphs — 20 domains. 4x F1, 11x fewer tokens, SHA-256 provenance. MCP-native.","url":"https://mcp.market/server/yarmoluk-ckg-nvidia-ai","rating":null,"grade":"B","score":70,"certified":false,"status":"active","category":"other","tags":[],"presence":{"score":16,"stars":null,"forks":null,"downloads_week":129,"last_push_at":null,"license":"MIT"},"uptime":{"percent":100,"checks":28,"ok":28,"last_checked_at":"2026-09-27T16:11:29.949Z","last_ok_at":"2026-09-27T16:11:29.949Z","latency_ms":466},"claimed":false,"transport":"mixed","callable_via_gateway":true,"default_price_micros":0,"repository":"https://github.com/Yarmoluk/ckg-nvidia-ai","website":null,"version":"0.16.4","remotes":[{"type":"streamable-http","url":"https://ckg-nvidia-ai.onrender.com/mcp"}],"packages":[{"registryType":"pypi","identifier":"ckg-nvidia-ai","version":"0.16.4","runtimeHint":"uvx","transport":{"type":"stdio"}}],"tools":[{"name":"ask_nvidia","description":"Ask a natural-language question answered by Qwen grounded on the NVIDIA CKG.\n\n    Requires Ollama running locally with a Qwen model pulled:\n        ollama pull qwen2.5:14b\n\n    Override model:  NVIDIA_CKG_MODEL env var  (default: qwen2.5:14b)\n    Override host:   NVIDIA_CKG_OLLAMA env var (default: http://localhost:11434)\n\n    Args:\n        question: Natural-language question about the NVIDIA AI stack.\n        domain:   Domain from list_domains() — auto-detected from question if omitted.\n    ","write_action":false,"price_micros":0,"input_schema":{"type":"object","properties":{"question":{"title":"Question","type":"string"},"domain":{"default":"","title":"Domain","type":"string"}},"required":["question"],"title":"ask_nvidiaArguments"}},{"name":"get_prerequisites","description":"Return the full ordered prerequisite chain for a concept — everything to learn first.\n\n    Args:\n        concept: Target concept — e.g. 'Speculative Decoding', 'Isaac Lab', 'NeMo Guardrails'.\n        domain:  Domain name from list_domains().\n    ","write_action":false,"price_micros":0,"input_schema":{"type":"object","properties":{"concept":{"title":"Concept","type":"string"},"domain":{"title":"Domain","type":"string"}},"required":["concept","domain"],"title":"get_prerequisitesArguments"}},{"name":"list_domains","description":"List all 20 NVIDIA AI domains available in this knowledge graph.","write_action":false,"price_micros":0,"input_schema":{"type":"object","properties":{},"title":"list_domainsArguments"}},{"name":"list_ecosystem","description":"Discover other CKG packages for adjacent domains — finance, healthcare, legal, and more.","write_action":false,"price_micros":0,"input_schema":{"type":"object","properties":{},"title":"list_ecosystemArguments"}},{"name":"query_ckg","description":"Traverse the NVIDIA knowledge graph from a concept — prerequisites and dependents.\n\n    Args:\n        concept: Concept name (partial match supported) — e.g. 'TensorRT', 'NIM', 'Isaac Lab'.\n        domain:  Domain name from list_domains() — e.g. 'nvidia-tensorrt-triton', 'nvidia-isaac'.\n        depth:   Traversal depth 1–5 (default 3).\n    ","write_action":false,"price_micros":0,"input_schema":{"type":"object","properties":{"concept":{"title":"Concept","type":"string"},"domain":{"title":"Domain","type":"string"},"depth":{"default":3,"title":"Depth","type":"integer"}},"required":["concept","domain"],"title":"query_ckgArguments"}},{"name":"query_intersect","description":"Answer a conjunctive query: concepts reachable from EVERY anchor at once (A AND B).\n\n    query_ckg walks outward from one concept. This intersects the reachable sets of two or\n    more, which is the shape of most real questions — \"the component that satisfies A AND\n    applies to B\". Neither anchor alone answers it; the answer lives in the overlap.\n\n    Every branch is an exact set of declared edges, so the intersection is exact. A concept\n    appears only if a declared path reaches it from each anchor. A relation missing from the\n    graph produces an empty result, never a guess.\n\n    Args:\n        branches: Two or more branches. Either a bare anchor (\"TensorRT-LLM\"), which takes\n            everything within `depth` hops, or an anchor plus an explicit relation path using\n            '>' (\"TensorRT-LLM > REQUIRES > ENABLES\"), where each relation replaces the\n            frontier. '*' matches any relation. Mix both forms freely.\n        domain: Domain name from list_domains().\n        depth: Hops for bare-anchor branches, 1-5 (default 2). Ignored for explicit paths.\n        direction: 'out' follows dependencies, 'in' follows them backwards, 'both' (default).\n        mode: 'AND' (default) intersects branches; 'OR' unions them.\n        limit: Max concepts listed, 1-200 (default 40). The true count is always shown.\n\n    Returns:\n        Markdown with the query plan and its per-step set sizes, then the answer set with\n        taxonomy tags. Reports which branch was empty when the intersection is empty.\n    ","write_action":false,"price_micros":0,"input_schema":{"type":"object","properties":{"branches":{"items":{"type":"string"},"title":"Branches","type":"array"},"domain":{"title":"Domain","type":"string"},"depth":{"default":2,"title":"Depth","type":"integer"},"direction":{"default":"both","title":"Direction","type":"string"},"mode":{"default":"AND","title":"Mode","type":"string"},"limit":{"default":40,"title":"Limit","type":"integer"}},"required":["branches","domain"],"title":"query_intersectArguments"}},{"name":"route_query","description":"Route an NVIDIA AI question to the optimal model and reasoning approach via graph depth.\n\n    The CKG graph IS the router — hop depth is a deterministic complexity metric.\n    Deeper NVIDIA prerequisite chains (CUDA → TensorRT → TensorRT-LLM → NIM) require\n    more capable models. No heuristic: the graph decides.\n\n    Routing table:\n      hop_depth 1  → haiku  · direct          (simple lookup)\n      hop_depth 2  → sonnet · generic_cot     (moderate chain)\n      hop_depth 3+ → opus   · sparql_cot      (deep dependency, structured reasoning)\n\n    Args:\n        question: Concept name or natural language question about NVIDIA AI.\n        domain:   Domain from list_domains() — e.g. \"nvidia-tensorrt-triton\", \"nvidia-nim\".\n\n    Returns:\n        model_tier + reasoning_approach + why + context subgraph to inject before LLM call.\n    ","write_action":false,"price_micros":0,"input_schema":{"type":"object","properties":{"question":{"title":"Question","type":"string"},"domain":{"default":"","title":"Domain","type":"string"}},"required":["question"],"title":"route_queryArguments"}},{"name":"search_concepts","description":"Find concepts in a NVIDIA AI domain by keyword.\n\n    Args:\n        query:  Search term — e.g. 'inference', 'sandbox', 'quantization', 'guardrails'.\n        domain: Domain name from list_domains() — e.g. 'nvidia-nim', 'nvidia-openshell'.\n    ","write_action":false,"price_micros":0,"input_schema":{"type":"object","properties":{"query":{"title":"Query","type":"string"},"domain":{"title":"Domain","type":"string"}},"required":["query","domain"],"title":"search_conceptsArguments"}},{"name":"verify_source","description":"Return the source URL and SHA-256 content hash for any NVIDIA AI concept node.\n\n    Audit chain: edge answer → graph commit → source_content_hash → source_url (fetch hint).\n    Verification: curl -s <source_url> | sha256sum  # compare to source_hash\n\n    Args:\n        concept: Concept label (partial match supported).\n        domain:  Domain from list_domains() — e.g. 'nvidia-nim', 'nvidia-tensorrt-triton'.\n    ","write_action":false,"price_micros":0,"input_schema":{"type":"object","properties":{"concept":{"title":"Concept","type":"string"},"domain":{"title":"Domain","type":"string"}},"required":["concept","domain"],"title":"verify_sourceArguments"}}],"scan":{"score":70,"grade":"B","scanned_at":"2026-09-24T17:25:17.271Z","report":{"scannerVersion":"0.1.9","scannedAt":"2026-09-24T17:25:17.162Z","components":{"code":{"score":20,"max":25,"notes":["54 source files scanned"]},"reliability":{"score":20,"max":20,"notes":["remote reachable in 973ms"]},"poisoning":{"score":15,"max":15,"notes":["9 tool descriptions checked"]},"auth":{"score":10,"max":15,"notes":["open endpoint, read-only tools"]},"maintenance":{"score":3,"max":15,"notes":["repository not readable: repo not found"]},"identity":{"score":2,"max":10,"notes":["no repository or website to verify"]}},"findings":[{"id":"exec.eval","severity":"medium","component":"code","title":"eval / new Function used","evidence":"ckg_nvidia_ai-0.16.4/site/assets/javascripts/lunr/wordcut.js: … __p;\\n'; try { var render = new Function(settings.variable || 'obj', '_', source)…"}],"inputs":{"probes":[{"url":"https://ckg-nvidia-ai.onrender.com/mcp","reachable":true,"authRequired":false,"latencyMs":973,"serverInfo":{"name":"ckg-nvidia-ai","version":"1.30.0"}}],"packages":[{"registryType":"pypi","identifier":"ckg-nvidia-ai","version":"0.16.4","found":true,"weeklyDownloads":129,"license":"MIT","dependencyCount":7,"publishedAt":"2026-07-29T16:57:24.753874Z","repositoryUrl":"https://github.com/Yarmoluk/ckg-nvidia-ai"}],"repo":{"found":false,"owner":"Yarmoluk","repo":"ckg-nvidia-ai","error":"repo not found"},"icon":{"url":null,"source":"none"},"presence":{"stars":null,"forks":null,"downloadsWeek":129,"license":"MIT","lastPushAt":null,"score":16}}}},"grade_history":[],"reviews":[]}