[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"skill:01M79FN1Q3C5MENKC908QNAYKB":3,"skill-same:01M79FN1Q3C5MENKC908QNAYKB":108},{"skillNo":4,"name":5,"nameZh":6,"description":7,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":16,"categories":17,"keywords":18,"publishedAt":19,"linkOnly":20,"markdownContent":21,"markdownBody":22,"frontMatter":23,"refName":48,"commitSha":49,"skillDirectory":50,"forksCount":14,"repositoryHtmlUrl":51,"repositoryHomepageUrl":6,"ownerHtmlUrl":52,"ownerBlogUrl":6,"ownerDescription":6,"updatedAt":53,"install":54,"resources":67},"01M79FN1Q3C5MENKC908QNAYKB","rag-eval",null,"Filesystem RAG benchmarks: corpus\u002F, train.json, evaluate_rag.py (RAGAS quality). Not for prod monitoring, latency\u002Fthroughput benchmarking (use rag-perf), or evals outside this repo layout.","Elogic360\u002Fsoftware-factory","Elogic360","https:\u002F\u002Favatars.githubusercontent.com\u002Fu\u002F156535519?v=4","01M74NH3F0RB13PWB7JZDE2V80","elogic360","01M74NH3FPC65082MDZVMFHDM3",0,"Apache-2.0","https:\u002F\u002Fgithub.com\u002FElogic360\u002Fsoftware-factory\u002Fblob\u002Fmain\u002Fskills\u002Fnvidia\u002Frag-eval\u002FSKILL.md",[],[],"2026-11-15 14:19:17",false,"---\nname: rag-eval\nversion: \"2.6.0\"\ndescription: >-\n  Filesystem RAG benchmarks: corpus\u002F, train.json, evaluate_rag.py (RAGAS quality). Not for prod\n  monitoring, latency\u002Fthroughput benchmarking (use rag-perf), or evals outside this repo layout.\nlicense: Apache-2.0\ncompatibility: Repository checkout with uv; Python 3.11+; run from repo root; uv sync --project scripts\u002Feval (eval deps live in scripts\u002Feval\u002Fpyproject.toml); network to RAG, ingestor, and vdb endpoints; NVIDIA_API_KEY for RAGAS; optional RAG_EVAL_JUDGE_MODEL (default mistralai\u002Fmixtral-8x22b-instruct-v0.1).\nmetadata:\n  author: NVIDIA RAG \u003Cfoundational-rag-dev@exchange.nvidia.com>\n  github-url: \"https:\u002F\u002Fgithub.com\u002FNVIDIA-AI-Blueprints\u002Frag\"\n  endpoint-openapi-schemas:\n    - docs\u002Fapi_reference\u002Fopenapi_schema_rag_server.json\n    - docs\u002Fapi_reference\u002Fopenapi_schema_ingestor_server.json\n  argument-hint: RAGAS eval | evaluate_rag | train.json | corpus | results json | error triage | uv run --project scripts\u002Feval | enable_reranker | query_rewriting | temperature | skip_ingestion\n  tags:\n    - nvidia\n    - blueprint\n    - rag\n    - evaluation\n    - ragas\n    - benchmarking\n    - nvidia-rag-blueprint\n  languages:\n    - python\n    - shell\n  frameworks:\n    - ragas\n    - fastapi\n  domain: ai-ml\nallowed-tools: Read Grep Glob Bash(ls *) Bash(python3 *) Bash(uv *) Write Edit\n---\n\n# On-disk RAG evaluation (`corpus\u002F` + `train.json`)\n\n## Purpose\n\nGuide agents through NVIDIA RAG Blueprint **filesystem** benchmarks: preparing `corpus\u002F` and `train.json`, running `scripts\u002Feval\u002Fevaluate_rag.py`, tuning retrieval and generation flags for **quality** comparisons, interpreting RAGAS JSON outputs, and triaging failures (HTTP\u002Fstream errors, empty contexts, collection mismatch, judge API).\n\nFor **latency, throughput, and load testing**, use the **rag-perf** skill (`scripts\u002Frag-perf`, `docs\u002Fperformance-benchmarking.md`) — not this skill.\n\n## When not to use\n\nDo **not** use this skill for: deploying or repairing services (use rag-blueprint); evaluating APIs without the `corpus\u002F` + `train.json` layout; general ML experimentation unrelated to this evaluator; production monitoring\u002Falerting; or latency\u002Fthroughput benchmarking (use **rag-perf**).\n\n## Prerequisites\n\n- Repo cloned; **run commands from repo root** (imports and paths assume this).\n- Python **3.11+** and **uv**; eval deps: `uv sync --project scripts\u002Feval`.\n- Reachable **RAG server** and **ingestor** (defaults often `localhost:8081` \u002F `8082`).\n- **`NVIDIA_API_KEY`** for RAGAS (see [credential hygiene](references\u002Fbenchmark-execution.md#credential-hygiene-nvidia_api_key)); optional **`RAG_EVAL_JUDGE_MODEL`**.\n- Dataset roots passed to `--dataset-paths` each contain **`corpus\u002F`** and **`train.json`**.\n\n## Instructions\n\n1. **Prepare data** — Ensure each dataset directory matches the layout and `train.json` rules in [`references\u002Fdataset-and-conversion.md`](references\u002Fdataset-and-conversion.md). When sources arrive as public links (sites or dataset pages), materialize documents under `corpus\u002F`—prefer **PDF** for multimodal content so **images stay embedded**; convert CSV\u002FJSONL\u002Fetc. using the patterns there.\n2. **Run eval** — `uv run --project scripts\u002Feval python scripts\u002Feval\u002Fevaluate_rag.py` with `--dataset-paths`, `--host`, and `--port`. See [`references\u002Fbenchmark-execution.md`](references\u002Fbenchmark-execution.md) for command examples, outputs, and errors. Use [`references\u002Fevaluate-rag-cli.md`](references\u002Fevaluate-rag-cli.md) for flag-level detail.\n3. **Tune quality** — Adjust `--top_k` \u002F `--vdb_top_k`, reranker and query-rewriting toggles, and generation overrides (`--temperature`, `--top-p`, `--max-tokens`) as documented in [`references\u002Fbenchmark-execution.md`](references\u002Fbenchmark-execution.md) when comparing retrieval\u002Fgeneration configs for RAGAS scores.\n4. **Analyze results** — Use [`references\u002Fresult-analysis.md`](references\u002Fresult-analysis.md) for scripts; scan `rag_*_evaluation_summary.json` for headline RAGAS metrics.\n5. **Triage errors** — Use the [error signal table](references\u002Fbenchmark-execution.md#common-error-cases-and-signals) and the **Troubleshooting** section below.\n\n## Examples\n\n**Set API key without putting secrets in shell history (preferred patterns):** load from a gitignored env file or secrets manager; avoid committing `.env`; rotate keys if exposed. Details: [`references\u002Fbenchmark-execution.md#credential-hygiene-nvidia_api_key`](references\u002Fbenchmark-execution.md#credential-hygiene-nvidia_api_key).\n\n**Minimal eval (key already in environment):**\n\n```bash\nuv sync --project scripts\u002Feval\nuv run --project scripts\u002Feval python scripts\u002Feval\u002Fevaluate_rag.py \\\n  --dataset-paths \u002Fpath\u002Fto\u002Fmy_dataset \\\n  --host localhost \\\n  --port 8081\n```\n\n**Pretty-print summary JSON:**\n\n```bash\npython3 -m json.tool results\u002Fmy_dataset\u002Frag_my_dataset_evaluation_summary.json\n```\n\nMore examples (skip ingestion, quality sweeps): [`references\u002Fbenchmark-execution.md`](references\u002Fbenchmark-execution.md).\n\n## Limitations\n\n- Evaluator behavior is fixed to the **filesystem contract** and `evaluate_rag.py`; it does not substitute for custom offline judges or non-RAG benchmarks.\n- **Vector DB \u002F embedding** choices follow deployed ingestor and RAG env — not overridden by this CLI alone.\n- **Scores depend on** retrieval quality, judge model availability, and `NVIDIA_API_KEY`; empty contexts yield partial RAGAS metrics (see references).\n- Large procedural detail lives under **`references\u002F`** to keep routing concise; read those files when the user needs step-by-step conversion, full flags, or error tables.\n\n## Troubleshooting\n\n| Error \u002F signal | Likely cause | What to do |\n|----------------|--------------|------------|\n| Immediate exit mentioning `NVIDIA_API_KEY` | Missing or invalid key | Set key via secure channel; see credential hygiene in [`references\u002Fbenchmark-execution.md`](references\u002Fbenchmark-execution.md). |\n| `train.json must be a JSON array` | Wrong JSON shape | Top-level array of objects; validate per [`references\u002Fdataset-and-conversion.md`](references\u002Fdataset-and-conversion.md). |\n| Fewer rows in `evaluation_data.json` than `train.json` | Per-query failures | Check stderr: network or stream JSON errors; see error table in benchmark-execution. |\n| Empty `generated_contexts` everywhere | Retrieval gap | Verify collection, ingestion, `top_k` \u002F `vdb_top_k`, and `ingestor_server_url` **without** `\u002Fv1` suffix. |\n| Ingestor 404 on upload | Bad ingestor base URL | Pass `http:\u002F\u002Fhost:port` only — code appends `\u002Fv1\u002F`. |\n\nFull signal table: [`references\u002Fbenchmark-execution.md#common-error-cases-and-signals`](references\u002Fbenchmark-execution.md#common-error-cases-and-signals).\n\n## Gotchas\n\n- **Run from repo root**: paths and imports in `scripts\u002Feval\u002Fevaluate_rag.py` assume this; a wrong directory silently breaks imports.\n- **`--ingestor_server_url`**: pass `http:\u002F\u002Fhost:port` without `\u002Fv1`—the code appends `\u002Fv1\u002F` automatically. Including `\u002Fv1` causes 404s on ingestor calls.\n- **Vector DB \u002F embedding settings**: not set by this CLI; configure via the deployed ingestor and RAG server env vars (e.g. `APP_VECTORSTORE_URL`, embedding model).\n- **`--model` \u002F `--llm_endpoint`**: forwarded verbatim only when explicitly set; omit to keep the server's configured LLM.\n- **Stale collections**: a previous run's ingested data persists unless you use `--force_ingestion`. Use `--collection` with a unique name when comparing quality across isolated runs.\n- **Empty context metrics**: if all `generated_contexts` are empty, RAGAS scores only `nv_accuracy` and leaves the other two metrics blank—this is not a silent success.\n\n## Source of truth\n\n| Piece | Location |\n|-------|----------|\n| Driver | `scripts\u002Feval\u002Fevaluate_rag.py` (`CORPUS_DIRECTORY` = `corpus`, `EVAL_DATA` = `train.json`) |\n| Human README (always in-repo) | `scripts\u002Feval\u002FREADME.md` |\n| Full CLI (flags, defaults) | `scripts\u002Feval\u002Fevaluate_rag.py --help`; [`references\u002Fevaluate-rag-cli.md`](references\u002Fevaluate-rag-cli.md) |\n| Dataset \u002F conversion | [`references\u002Fdataset-and-conversion.md`](references\u002Fdataset-and-conversion.md) |\n| Runs, outputs, errors | [`references\u002Fbenchmark-execution.md`](references\u002Fbenchmark-execution.md) |\n| Result analysis scripts | [`references\u002Fresult-analysis.md`](references\u002Fresult-analysis.md) |\n| Latency \u002F throughput | **rag-perf** skill, `docs\u002Fperformance-benchmarking.md` |\n\n## Agent playbook\n\n1. **Run eval** — `uv sync --project scripts\u002Feval` then `uv run --project scripts\u002Feval python scripts\u002Feval\u002Fevaluate_rag.py` with required `--dataset-paths`, `--host`, and `--port` (and env `NVIDIA_API_KEY`). Argument `--ingestor_server_url` is optional (defaults to `http:\u002F\u002Flocalhost:8082`); pass it only when overriding the ingestor endpoint.\n2. **Quality tuning** — See [`references\u002Fbenchmark-execution.md`](references\u002Fbenchmark-execution.md): `--top_k`\u002F`--vdb_top_k`, reranker and query-rewriting toggles, `--temperature`, `--top-p`, `--max-tokens`.\n3. **Data conversion** — Follow [`references\u002Fdataset-and-conversion.md`](references\u002Fdataset-and-conversion.md).\n4. **Analyze results** — [`references\u002Fresult-analysis.md`](references\u002Fresult-analysis.md); quick scan: `python3 -m json.tool results\u002F\u003Cdataset>\u002Frag_\u003Cdataset>_evaluation_summary.json`.\n5. **Error triage** — [`references\u002Fbenchmark-execution.md#common-error-cases-and-signals`](references\u002Fbenchmark-execution.md#common-error-cases-and-signals).\n","# On-disk RAG evaluation (`corpus\u002F` + `train.json`)\n\n## Purpose\n\nGuide agents through NVIDIA RAG Blueprint **filesystem** benchmarks: preparing `corpus\u002F` and `train.json`, running `scripts\u002Feval\u002Fevaluate_rag.py`, tuning retrieval and generation flags for **quality** comparisons, interpreting RAGAS JSON outputs, and triaging failures (HTTP\u002Fstream errors, empty contexts, collection mismatch, judge API).\n\nFor **latency, throughput, and load testing**, use the **rag-perf** skill (`scripts\u002Frag-perf`, `docs\u002Fperformance-benchmarking.md`) — not this skill.\n\n## When not to use\n\nDo **not** use this skill for: deploying or repairing services (use rag-blueprint); evaluating APIs without the `corpus\u002F` + `train.json` layout; general ML experimentation unrelated to this evaluator; production monitoring\u002Falerting; or latency\u002Fthroughput benchmarking (use **rag-perf**).\n\n## Prerequisites\n\n- Repo cloned; **run commands from repo root** (imports and paths assume this).\n- Python **3.11+** and **uv**; eval deps: `uv sync --project scripts\u002Feval`.\n- Reachable **RAG server** and **ingestor** (defaults often `localhost:8081` \u002F `8082`).\n- **`NVIDIA_API_KEY`** for RAGAS (see [credential hygiene](references\u002Fbenchmark-execution.md#credential-hygiene-nvidia_api_key)); optional **`RAG_EVAL_JUDGE_MODEL`**.\n- Dataset roots passed to `--dataset-paths` each contain **`corpus\u002F`** and **`train.json`**.\n\n## Instructions\n\n1. **Prepare data** — Ensure each dataset directory matches the layout and `train.json` rules in [`references\u002Fdataset-and-conversion.md`](references\u002Fdataset-and-conversion.md). When sources arrive as public links (sites or dataset pages), materialize documents under `corpus\u002F`—prefer **PDF** for multimodal content so **images stay embedded**; convert CSV\u002FJSONL\u002Fetc. using the patterns there.\n2. **Run eval** — `uv run --project scripts\u002Feval python scripts\u002Feval\u002Fevaluate_rag.py` with `--dataset-paths`, `--host`, and `--port`. See [`references\u002Fbenchmark-execution.md`](references\u002Fbenchmark-execution.md) for command examples, outputs, and errors. Use [`references\u002Fevaluate-rag-cli.md`](references\u002Fevaluate-rag-cli.md) for flag-level detail.\n3. **Tune quality** — Adjust `--top_k` \u002F `--vdb_top_k`, reranker and query-rewriting toggles, and generation overrides (`--temperature`, `--top-p`, `--max-tokens`) as documented in [`references\u002Fbenchmark-execution.md`](references\u002Fbenchmark-execution.md) when comparing retrieval\u002Fgeneration configs for RAGAS scores.\n4. **Analyze results** — Use [`references\u002Fresult-analysis.md`](references\u002Fresult-analysis.md) for scripts; scan `rag_*_evaluation_summary.json` for headline RAGAS metrics.\n5. **Triage errors** — Use the [error signal table](references\u002Fbenchmark-execution.md#common-error-cases-and-signals) and the **Troubleshooting** section below.\n\n## Examples\n\n**Set API key without putting secrets in shell history (preferred patterns):** load from a gitignored env file or secrets manager; avoid committing `.env`; rotate keys if exposed. Details: [`references\u002Fbenchmark-execution.md#credential-hygiene-nvidia_api_key`](references\u002Fbenchmark-execution.md#credential-hygiene-nvidia_api_key).\n\n**Minimal eval (key already in environment):**\n\n```bash\nuv sync --project scripts\u002Feval\nuv run --project scripts\u002Feval python scripts\u002Feval\u002Fevaluate_rag.py \\\n  --dataset-paths \u002Fpath\u002Fto\u002Fmy_dataset \\\n  --host localhost \\\n  --port 8081\n```\n\n**Pretty-print summary JSON:**\n\n```bash\npython3 -m json.tool results\u002Fmy_dataset\u002Frag_my_dataset_evaluation_summary.json\n```\n\nMore examples (skip ingestion, quality sweeps): [`references\u002Fbenchmark-execution.md`](references\u002Fbenchmark-execution.md).\n\n## Limitations\n\n- Evaluator behavior is fixed to the **filesystem contract** and `evaluate_rag.py`; it does not substitute for custom offline judges or non-RAG benchmarks.\n- **Vector DB \u002F embedding** choices follow deployed ingestor and RAG env — not overridden by this CLI alone.\n- **Scores depend on** retrieval quality, judge model availability, and `NVIDIA_API_KEY`; empty contexts yield partial RAGAS metrics (see references).\n- Large procedural detail lives under **`references\u002F`** to keep routing concise; read those files when the user needs step-by-step conversion, full flags, or error tables.\n\n## Troubleshooting\n\n| Error \u002F signal | Likely cause | What to do |\n|----------------|--------------|------------|\n| Immediate exit mentioning `NVIDIA_API_KEY` | Missing or invalid key | Set key via secure channel; see credential hygiene in [`references\u002Fbenchmark-execution.md`](references\u002Fbenchmark-execution.md). |\n| `train.json must be a JSON array` | Wrong JSON shape | Top-level array of objects; validate per [`references\u002Fdataset-and-conversion.md`](references\u002Fdataset-and-conversion.md). |\n| Fewer rows in `evaluation_data.json` than `train.json` | Per-query failures | Check stderr: network or stream JSON errors; see error table in benchmark-execution. |\n| Empty `generated_contexts` everywhere | Retrieval gap | Verify collection, ingestion, `top_k` \u002F `vdb_top_k`, and `ingestor_server_url` **without** `\u002Fv1` suffix. |\n| Ingestor 404 on upload | Bad ingestor base URL | Pass `http:\u002F\u002Fhost:port` only — code appends `\u002Fv1\u002F`. |\n\nFull signal table: [`references\u002Fbenchmark-execution.md#common-error-cases-and-signals`](references\u002Fbenchmark-execution.md#common-error-cases-and-signals).\n\n## Gotchas\n\n- **Run from repo root**: paths and imports in `scripts\u002Feval\u002Fevaluate_rag.py` assume this; a wrong directory silently breaks imports.\n- **`--ingestor_server_url`**: pass `http:\u002F\u002Fhost:port` without `\u002Fv1`—the code appends `\u002Fv1\u002F` automatically. Including `\u002Fv1` causes 404s on ingestor calls.\n- **Vector DB \u002F embedding settings**: not set by this CLI; configure via the deployed ingestor and RAG server env vars (e.g. `APP_VECTORSTORE_URL`, embedding model).\n- **`--model` \u002F `--llm_endpoint`**: forwarded verbatim only when explicitly set; omit to keep the server's configured LLM.\n- **Stale collections**: a previous run's ingested data persists unless you use `--force_ingestion`. Use `--collection` with a unique name when comparing quality across isolated runs.\n- **Empty context metrics**: if all `generated_contexts` are empty, RAGAS scores only `nv_accuracy` and leaves the other two metrics blank—this is not a silent success.\n\n## Source of truth\n\n| Piece | Location |\n|-------|----------|\n| Driver | `scripts\u002Feval\u002Fevaluate_rag.py` (`CORPUS_DIRECTORY` = `corpus`, `EVAL_DATA` = `train.json`) |\n| Human README (always in-repo) | `scripts\u002Feval\u002FREADME.md` |\n| Full CLI (flags, defaults) | `scripts\u002Feval\u002Fevaluate_rag.py --help`; [`references\u002Fevaluate-rag-cli.md`](references\u002Fevaluate-rag-cli.md) |\n| Dataset \u002F conversion | [`references\u002Fdataset-and-conversion.md`](references\u002Fdataset-and-conversion.md) |\n| Runs, outputs, errors | [`references\u002Fbenchmark-execution.md`](references\u002Fbenchmark-execution.md) |\n| Result analysis scripts | [`references\u002Fresult-analysis.md`](references\u002Fresult-analysis.md) |\n| Latency \u002F throughput | **rag-perf** skill, `docs\u002Fperformance-benchmarking.md` |\n\n## Agent playbook\n\n1. **Run eval** — `uv sync --project scripts\u002Feval` then `uv run --project scripts\u002Feval python scripts\u002Feval\u002Fevaluate_rag.py` with required `--dataset-paths`, `--host`, and `--port` (and env `NVIDIA_API_KEY`). Argument `--ingestor_server_url` is optional (defaults to `http:\u002F\u002Flocalhost:8082`); pass it only when overriding the ingestor endpoint.\n2. **Quality tuning** — See [`references\u002Fbenchmark-execution.md`](references\u002Fbenchmark-execution.md): `--top_k`\u002F`--vdb_top_k`, reranker and query-rewriting toggles, `--temperature`, `--top-p`, `--max-tokens`.\n3. **Data conversion** — Follow [`references\u002Fdataset-and-conversion.md`](references\u002Fdataset-and-conversion.md).\n4. **Analyze results** — [`references\u002Fresult-analysis.md`](references\u002Fresult-analysis.md); quick scan: `python3 -m json.tool results\u002F\u003Cdataset>\u002Frag_\u003Cdataset>_evaluation_summary.json`.\n5. **Error triage** — [`references\u002Fbenchmark-execution.md#common-error-cases-and-signals`](references\u002Fbenchmark-execution.md#common-error-cases-and-signals).\n",{"name":5,"license":15,"version":24,"metadata":25,"description":7,"allowed-tools":46,"compatibility":47},"2.6.0",{"tags":26,"author":34,"domain":35,"languages":36,"frameworks":39,"github-url":41,"argument-hint":42,"endpoint-openapi-schemas":43},[27,28,29,30,31,32,33],"nvidia","blueprint","rag","evaluation","ragas","benchmarking","nvidia-rag-blueprint","NVIDIA RAG \u003Cfoundational-rag-dev@exchange.nvidia.com>","ai-ml",[37,38],"python","shell",[31,40],"fastapi","https:\u002F\u002Fgithub.com\u002FNVIDIA-AI-Blueprints\u002Frag","RAGAS eval | evaluate_rag | train.json | corpus | results json | error triage | uv run --project scripts\u002Feval | enable_reranker | query_rewriting | temperature | skip_ingestion",[44,45],"docs\u002Fapi_reference\u002Fopenapi_schema_rag_server.json","docs\u002Fapi_reference\u002Fopenapi_schema_ingestor_server.json","Read Grep Glob Bash(ls *) Bash(python3 *) Bash(uv *) Write Edit","Repository checkout with uv; Python 3.11+; run from repo root; uv sync --project scripts\u002Feval (eval deps live in scripts\u002Feval\u002Fpyproject.toml); network to RAG, ingestor, and vdb endpoints; NVIDIA_API_KEY for RAGAS; optional RAG_EVAL_JUDGE_MODEL (default mistralai\u002Fmixtral-8x22b-instruct-v0.1).","main","c475fc64e94e803b50d14a791dfee8853fbb2c5c","skills\u002Fnvidia\u002Frag-eval","https:\u002F\u002Fgithub.com\u002FElogic360\u002Fsoftware-factory","https:\u002F\u002Fgithub.com\u002FElogic360","2026-11-08 22:21:22",{"source":8,"skillName":5,"commands":55,"prompt":65,"zipAvailable":66},[56,59,62],{"tool":57,"command":58},"npx","npx skills add Elogic360\u002Fsoftware-factory --skill rag-eval",{"tool":60,"command":61},"bunx","bunx skills add Elogic360\u002Fsoftware-factory --skill rag-eval",{"tool":63,"command":64},"pnpm","pnpm dlx skills add Elogic360\u002Fsoftware-factory --skill rag-eval","安装技能，技能地址：https:\u002F\u002Fgithub.com\u002FElogic360\u002Fsoftware-factory\u002Fblob\u002Fmain\u002Fskills\u002Fnvidia\u002Frag-eval\u002FSKILL.md",true,[68,74,78,82,86,90,94,98,102],{"path":69,"resourceType":70,"resourceTypeDesc":71,"sizeBytes":72,"sourceHtmlUrl":73},"BENCHMARK.md","REFERENCE","参考文档",2915,"https:\u002F\u002Fgithub.com\u002FElogic360\u002Fsoftware-factory\u002Fblob\u002Fmain\u002Fskills\u002Fnvidia\u002Frag-eval\u002FBENCHMARK.md",{"path":75,"resourceType":70,"resourceTypeDesc":71,"sizeBytes":76,"sourceHtmlUrl":77},"eval\u002Fh100.json",2893,"https:\u002F\u002Fgithub.com\u002FElogic360\u002Fsoftware-factory\u002Fblob\u002Fmain\u002Fskills\u002Fnvidia\u002Frag-eval\u002Feval\u002Fh100.json",{"path":79,"resourceType":70,"resourceTypeDesc":71,"sizeBytes":80,"sourceHtmlUrl":81},"eval\u002Fnvidia_hosted.json",1963,"https:\u002F\u002Fgithub.com\u002FElogic360\u002Fsoftware-factory\u002Fblob\u002Fmain\u002Fskills\u002Fnvidia\u002Frag-eval\u002Feval\u002Fnvidia_hosted.json",{"path":83,"resourceType":70,"resourceTypeDesc":71,"sizeBytes":84,"sourceHtmlUrl":85},"references\u002Fbenchmark-execution.md",8315,"https:\u002F\u002Fgithub.com\u002FElogic360\u002Fsoftware-factory\u002Fblob\u002Fmain\u002Fskills\u002Fnvidia\u002Frag-eval\u002Freferences\u002Fbenchmark-execution.md",{"path":87,"resourceType":70,"resourceTypeDesc":71,"sizeBytes":88,"sourceHtmlUrl":89},"references\u002Fdataset-and-conversion.md",8361,"https:\u002F\u002Fgithub.com\u002FElogic360\u002Fsoftware-factory\u002Fblob\u002Fmain\u002Fskills\u002Fnvidia\u002Frag-eval\u002Freferences\u002Fdataset-and-conversion.md",{"path":91,"resourceType":70,"resourceTypeDesc":71,"sizeBytes":92,"sourceHtmlUrl":93},"references\u002Fevaluate-rag-cli.md",4283,"https:\u002F\u002Fgithub.com\u002FElogic360\u002Fsoftware-factory\u002Fblob\u002Fmain\u002Fskills\u002Fnvidia\u002Frag-eval\u002Freferences\u002Fevaluate-rag-cli.md",{"path":95,"resourceType":70,"resourceTypeDesc":71,"sizeBytes":96,"sourceHtmlUrl":97},"references\u002Fresult-analysis.md",2868,"https:\u002F\u002Fgithub.com\u002FElogic360\u002Fsoftware-factory\u002Fblob\u002Fmain\u002Fskills\u002Fnvidia\u002Frag-eval\u002Freferences\u002Fresult-analysis.md",{"path":99,"resourceType":70,"resourceTypeDesc":71,"sizeBytes":100,"sourceHtmlUrl":101},"skill-card.md",2673,"https:\u002F\u002Fgithub.com\u002FElogic360\u002Fsoftware-factory\u002Fblob\u002Fmain\u002Fskills\u002Fnvidia\u002Frag-eval\u002Fskill-card.md",{"path":103,"resourceType":104,"resourceTypeDesc":105,"sizeBytes":106,"sourceHtmlUrl":107},"skill.oms.sig","OTHER","其他",5797,"https:\u002F\u002Fgithub.com\u002FElogic360\u002Fsoftware-factory\u002Fblob\u002Fmain\u002Fskills\u002Fnvidia\u002Frag-eval\u002Fskill.oms.sig",[109,116,123,130,137,144],{"skillNo":110,"name":111,"nameZh":6,"description":112,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":113,"categories":114,"keywords":115,"publishedAt":19},"01M79G206MT3J6S0GY3C45P2KV","vss-summarize-video","Use to summarize a recorded video via the LVS summarization microservice (HITL-gated) with a VLM fallback. Not for report generation or live RTSP captioning.","https:\u002F\u002Fgithub.com\u002FElogic360\u002Fsoftware-factory\u002Fblob\u002Fmain\u002Fskills\u002Fnvidia\u002Fvss-summarize-video\u002FSKILL.md",[],[],{"skillNo":117,"name":118,"nameZh":6,"description":119,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":120,"categories":121,"keywords":122,"publishedAt":19},"01M79G1WN63PGRE8CKD47WEBQ5","vss-setup-video-analytics-api","Use to deploy the vss-video-analytics-api REST service standalone (config-source, data-log bind, Elasticsearch, optional Kafka). Not for full warehouse deploy.","https:\u002F\u002Fgithub.com\u002FElogic360\u002Fsoftware-factory\u002Fblob\u002Fmain\u002Fskills\u002Fnvidia\u002Fvss-setup-video-analytics-api\u002FSKILL.md",[],[],{"skillNo":124,"name":125,"nameZh":6,"description":126,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":127,"categories":128,"keywords":129,"publishedAt":19},"01M79G1S9F7VY3EF6H53PRP5AN","vss-setup-behavior-analytics","Use to deploy the vss-behavior-analytics service standalone (entrypoint, config-source, optional calibration). Not for the full warehouse deploy.","https:\u002F\u002Fgithub.com\u002FElogic360\u002Fsoftware-factory\u002Fblob\u002Fmain\u002Fskills\u002Fnvidia\u002Fvss-setup-behavior-analytics\u002FSKILL.md",[],[],{"skillNo":131,"name":132,"nameZh":6,"description":133,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":134,"categories":135,"keywords":136,"publishedAt":19},"01M79G1N1S1Y27PFZPYMNK1MTQ","vss-search-archive","Use this skill to run top-level VSS fusion search on archived video, or to ingest video files \u002F RTSP streams for search. Do NOT use for ad-hoc visual Q&A (use vss-ask-video), live captioning (use vss-deploy-dense-captioning), or video summarization and reports (use vss-summarize-video).","https:\u002F\u002Fgithub.com\u002FElogic360\u002Fsoftware-factory\u002Fblob\u002Fmain\u002Fskills\u002Fnvidia\u002Fvss-search-archive\u002FSKILL.md",[],[],{"skillNo":138,"name":139,"nameZh":6,"description":140,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":141,"categories":142,"keywords":143,"publishedAt":19},"01M79G1H6WN0Q6C0VCAXCZX6B7","vss-query-analytics","Use this skill when reading video-analytics metrics, incidents, alerts, and sensor data via the VA-MCP server (port 9901). Not for live VLM or incident-range narrative reports.","https:\u002F\u002Fgithub.com\u002FElogic360\u002Fsoftware-factory\u002Fblob\u002Fmain\u002Fskills\u002Fnvidia\u002Fvss-query-analytics\u002FSKILL.md",[],[],{"skillNo":145,"name":146,"nameZh":6,"description":147,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":148,"categories":149,"keywords":150,"publishedAt":19},"01M79G1DYPNXQS7QHZ0AAF5FX5","vss-manage-video-io-storage","Use to call the VIOS REST API (sensor list, timelines, clip extraction, snapshots, add\u002Fdelete sensors and streams). Not for VLM inference or search.","https:\u002F\u002Fgithub.com\u002FElogic360\u002Fsoftware-factory\u002Fblob\u002Fmain\u002Fskills\u002Fnvidia\u002Fvss-manage-video-io-storage\u002FSKILL.md",[],[]]