[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"skill:01M7ERD9SSH3HC7EJR88R522J4":3,"skill-same:01M7ERD9SSH3HC7EJR88R522J4":48},{"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":26,"commitSha":27,"skillDirectory":28,"forksCount":29,"repositoryHtmlUrl":30,"repositoryHomepageUrl":6,"ownerHtmlUrl":31,"ownerBlogUrl":6,"ownerDescription":32,"updatedAt":33,"install":34,"resources":47},"01M7ERD9SSH3HC7EJR88R522J4","vector-search",null,"Vector search via embeddings_* (large-scale HNSW) and ruvllm_hnsw_* (WASM router for ≤11 hot patterns), with RaBitQ 1-bit quantization for 32× memory reduction","lza6\u002FArmor-piercing-system-prompts","lza6","https:\u002F\u002Favatars.githubusercontent.com\u002Fu\u002F128880206?v=4","01M7EB3K7WPX5T28EJ2NVHE6EG","Chinese-tingfeng","01M7EB3K97E1KQ16QWKJMFNGX8",11,"MIT","https:\u002F\u002Fgithub.com\u002Flza6\u002FArmor-piercing-system-prompts\u002Fblob\u002Fmain\u002F04-安全渗透\u002Fruflo\u002Fplugins\u002Fruflo-agentdb\u002Fskills\u002Fvector-search\u002FSKILL.md",[],[],"2026-11-15 14:19:17",false,"---\nname: vector-search\ndescription: Vector search via embeddings_* (large-scale HNSW) and ruvllm_hnsw_* (WASM router for ≤11 hot patterns), with RaBitQ 1-bit quantization for 32× memory reduction\nargument-hint: \"\u003Cquery> [--limit N] [--quantized]\"\nallowed-tools: mcp__plugin_ruflo-core_ruflo__embeddings_generate mcp__plugin_ruflo-core_ruflo__embeddings_search mcp__plugin_ruflo-core_ruflo__embeddings_compare mcp__plugin_ruflo-core_ruflo__embeddings_init mcp__plugin_ruflo-core_ruflo__embeddings_status mcp__plugin_ruflo-core_ruflo__embeddings_hyperbolic mcp__plugin_ruflo-core_ruflo__embeddings_neural mcp__plugin_ruflo-core_ruflo__embeddings_rabitq_build mcp__plugin_ruflo-core_ruflo__embeddings_rabitq_search mcp__plugin_ruflo-core_ruflo__embeddings_rabitq_status mcp__plugin_ruflo-core_ruflo__ruvllm_hnsw_create mcp__plugin_ruflo-core_ruflo__ruvllm_hnsw_add mcp__plugin_ruflo-core_ruflo__ruvllm_hnsw_route mcp__plugin_ruflo-core_ruflo__memory_search_unified Bash\n---\n\n# Vector Search\n\nTwo distinct vector-search paths live in this plugin. Pick the right one — they're not interchangeable.\n\n| Path | Tool family | Backing | Capacity | Latency |\n|------|-------------|---------|----------|---------|\n| **Large-scale corpus** | `embeddings_*` | `@claude-flow\u002Fmemory` HNSW (Rust\u002FNative) | up to millions of vectors | ~1.9× at N=20k, ~3.2×–4.7× at N=5k vs brute-force (measured; recall@10 ≈ 0.99). ANN wins above the crossover |\n| **Hot-path router** | `ruvllm_hnsw_*` | WASM-backed router (v2.0.1) | **~11 patterns max** (`ruvllm-tools.ts:58`) | sub-ms; designed for high-priority routing, not corpus search |\n\nThe \"12,500×\" headline applies to the large-scale `embeddings_search` path. The WASM router is **not** that path.\n\n## When to use\n\n| Need | Path |\n|---|---|\n| Search a corpus of N ≥ 500 documents | `embeddings_search` |\n| Memory-constrained corpus (≥5,000 vectors) | RaBitQ quantized — see \"Quantized search\" below |\n| Compare two strings | `embeddings_compare` |\n| Hierarchical \u002F taxonomic data | `embeddings_hyperbolic` (Poincare ball) |\n| Route a query to one of ≤11 hot patterns | `ruvllm_hnsw_route` |\n| Cross-namespace search | `memory_search_unified` |\n\n## Standard search\n\n1. **Check status** — `mcp__plugin_ruflo-core_ruflo__embeddings_status` to verify the embedding engine.\n2. **Initialize** — `mcp__plugin_ruflo-core_ruflo__embeddings_init` if not active.\n3. **Generate** — `mcp__plugin_ruflo-core_ruflo__embeddings_generate` for text input.\n4. **Search** — `mcp__plugin_ruflo-core_ruflo__embeddings_search` with the query.\n5. **Compare** — `mcp__plugin_ruflo-core_ruflo__embeddings_compare` to measure similarity.\n6. **Unified search** — `mcp__plugin_ruflo-core_ruflo__memory_search_unified` for cross-namespace.\n\n## Quantized search (32× memory reduction)\n\nFor corpora ≥5,000 vectors and\u002For memory-constrained environments, use the RaBitQ 1-bit quantization workflow. Below 5,000 vectors the rebuild cost outweighs the savings — use the standard path instead.\n\n| Step | Tool | Purpose |\n|---|---|---|\n| 1 | `embeddings_init` | Engine warm |\n| 2 | `embeddings_rabitq_build` | One-time build of the 1-bit index after corpus is loaded |\n| 3 | `embeddings_rabitq_search` | Hamming-prefilter returns top-N candidate IDs (cheap) |\n| 4 | `embeddings_search` | Optional exact rerank on the candidate set (full-precision) |\n| 5 | `embeddings_rabitq_status` | Index health, memory footprint, build time |\n\n> **Note**: `embeddings_rabitq_search` returns candidate IDs only — the rerank in step 4 is the user's responsibility (mirrors the docstring at `embeddings-tools.ts:911`). Without rerank, results are approximate; with rerank, you get full-precision quality at 32× lower memory.\n\n## Tuning\n\nHNSW exposes three knobs that trade recall against latency. The \"12,500×\" headline assumes **defaults**; tune deliberately for your workload:\n\n| Profile | `efSearch` | `M` | When to use |\n|---------|-----------|-----|-------------|\n| `recall-first` | 200 | 32 | Pattern recall during planning; quality matters more than ms |\n| `balanced` (default) | 64 | 16 | General-purpose semantic recall |\n| `latency-first` | 16 | 8 | Hot-path routing where p99 latency matters |\n\n`efSearch` is passed via `ruvllm_hnsw_create` (`ruvllm-tools.ts:64`). `M` is registry-level today; raise as a follow-up if it should be MCP-tunable. `efConstruction` defaults to 200 in the lite index (`hnsw-index.ts:537`).\n\n## HNSW pattern router (WASM, ≤11 patterns)\n\nFor routing a small number of high-priority patterns:\n- `mcp__plugin_ruflo-core_ruflo__ruvllm_hnsw_create` — create the WASM index (cap ~11)\n- `mcp__plugin_ruflo-core_ruflo__ruvllm_hnsw_add` — add a pattern\n- `mcp__plugin_ruflo-core_ruflo__ruvllm_hnsw_route` — route an incoming query\n\nThis is **not** a corpus index. Treat it as a fast classifier over a curated set of patterns.\n\n## Hyperbolic embeddings\n\nFor hierarchical data (code trees, org charts), use `mcp__plugin_ruflo-core_ruflo__embeddings_hyperbolic` which maps to Poincare ball space. Distance is geodesic, not cosine.\n\n## CLI alternative\n\n```bash\nnpx @claude-flow\u002Fcli@latest embeddings search --query \"authentication patterns\"\nnpx @claude-flow\u002Fcli@latest embeddings init\nnpx @claude-flow\u002Fcli@latest memory search --query \"your query\"\n```\n\n## Performance\n\nMeasured numbers (source: `scripts\u002Fbenchmark-intelligence.mjs`, ruvector NAPI backend; recall@10 ≈ 0.99). The older \"150×–12,500×\" figures were brute-force-fallback artifacts and have been retired — see project CLAUDE.md \"V3 Performance Targets\".\n\n| Method | Measured speedup vs brute-force |\n|--------|---------------------------------|\n| Brute-force scan | Baseline |\n| HNSW (N=5,000) | ~3.2×–4.7× faster |\n| HNSW (N=20,000) | ~1.9× faster |\n| HNSW (below crossover, small N) | ties\u002Floses vs brute-force |\n| RaBitQ quantization | 32× memory reduction; 0.60 ms\u002Fquery at N≈14.7k |\n| `ruvllm_hnsw_route` (n≤11) | sub-ms per route, fixed cost |\n","# Vector Search\n\nTwo distinct vector-search paths live in this plugin. Pick the right one — they're not interchangeable.\n\n| Path | Tool family | Backing | Capacity | Latency |\n|------|-------------|---------|----------|---------|\n| **Large-scale corpus** | `embeddings_*` | `@claude-flow\u002Fmemory` HNSW (Rust\u002FNative) | up to millions of vectors | ~1.9× at N=20k, ~3.2×–4.7× at N=5k vs brute-force (measured; recall@10 ≈ 0.99). ANN wins above the crossover |\n| **Hot-path router** | `ruvllm_hnsw_*` | WASM-backed router (v2.0.1) | **~11 patterns max** (`ruvllm-tools.ts:58`) | sub-ms; designed for high-priority routing, not corpus search |\n\nThe \"12,500×\" headline applies to the large-scale `embeddings_search` path. The WASM router is **not** that path.\n\n## When to use\n\n| Need | Path |\n|---|---|\n| Search a corpus of N ≥ 500 documents | `embeddings_search` |\n| Memory-constrained corpus (≥5,000 vectors) | RaBitQ quantized — see \"Quantized search\" below |\n| Compare two strings | `embeddings_compare` |\n| Hierarchical \u002F taxonomic data | `embeddings_hyperbolic` (Poincare ball) |\n| Route a query to one of ≤11 hot patterns | `ruvllm_hnsw_route` |\n| Cross-namespace search | `memory_search_unified` |\n\n## Standard search\n\n1. **Check status** — `mcp__plugin_ruflo-core_ruflo__embeddings_status` to verify the embedding engine.\n2. **Initialize** — `mcp__plugin_ruflo-core_ruflo__embeddings_init` if not active.\n3. **Generate** — `mcp__plugin_ruflo-core_ruflo__embeddings_generate` for text input.\n4. **Search** — `mcp__plugin_ruflo-core_ruflo__embeddings_search` with the query.\n5. **Compare** — `mcp__plugin_ruflo-core_ruflo__embeddings_compare` to measure similarity.\n6. **Unified search** — `mcp__plugin_ruflo-core_ruflo__memory_search_unified` for cross-namespace.\n\n## Quantized search (32× memory reduction)\n\nFor corpora ≥5,000 vectors and\u002For memory-constrained environments, use the RaBitQ 1-bit quantization workflow. Below 5,000 vectors the rebuild cost outweighs the savings — use the standard path instead.\n\n| Step | Tool | Purpose |\n|---|---|---|\n| 1 | `embeddings_init` | Engine warm |\n| 2 | `embeddings_rabitq_build` | One-time build of the 1-bit index after corpus is loaded |\n| 3 | `embeddings_rabitq_search` | Hamming-prefilter returns top-N candidate IDs (cheap) |\n| 4 | `embeddings_search` | Optional exact rerank on the candidate set (full-precision) |\n| 5 | `embeddings_rabitq_status` | Index health, memory footprint, build time |\n\n> **Note**: `embeddings_rabitq_search` returns candidate IDs only — the rerank in step 4 is the user's responsibility (mirrors the docstring at `embeddings-tools.ts:911`). Without rerank, results are approximate; with rerank, you get full-precision quality at 32× lower memory.\n\n## Tuning\n\nHNSW exposes three knobs that trade recall against latency. The \"12,500×\" headline assumes **defaults**; tune deliberately for your workload:\n\n| Profile | `efSearch` | `M` | When to use |\n|---------|-----------|-----|-------------|\n| `recall-first` | 200 | 32 | Pattern recall during planning; quality matters more than ms |\n| `balanced` (default) | 64 | 16 | General-purpose semantic recall |\n| `latency-first` | 16 | 8 | Hot-path routing where p99 latency matters |\n\n`efSearch` is passed via `ruvllm_hnsw_create` (`ruvllm-tools.ts:64`). `M` is registry-level today; raise as a follow-up if it should be MCP-tunable. `efConstruction` defaults to 200 in the lite index (`hnsw-index.ts:537`).\n\n## HNSW pattern router (WASM, ≤11 patterns)\n\nFor routing a small number of high-priority patterns:\n- `mcp__plugin_ruflo-core_ruflo__ruvllm_hnsw_create` — create the WASM index (cap ~11)\n- `mcp__plugin_ruflo-core_ruflo__ruvllm_hnsw_add` — add a pattern\n- `mcp__plugin_ruflo-core_ruflo__ruvllm_hnsw_route` — route an incoming query\n\nThis is **not** a corpus index. Treat it as a fast classifier over a curated set of patterns.\n\n## Hyperbolic embeddings\n\nFor hierarchical data (code trees, org charts), use `mcp__plugin_ruflo-core_ruflo__embeddings_hyperbolic` which maps to Poincare ball space. Distance is geodesic, not cosine.\n\n## CLI alternative\n\n```bash\nnpx @claude-flow\u002Fcli@latest embeddings search --query \"authentication patterns\"\nnpx @claude-flow\u002Fcli@latest embeddings init\nnpx @claude-flow\u002Fcli@latest memory search --query \"your query\"\n```\n\n## Performance\n\nMeasured numbers (source: `scripts\u002Fbenchmark-intelligence.mjs`, ruvector NAPI backend; recall@10 ≈ 0.99). The older \"150×–12,500×\" figures were brute-force-fallback artifacts and have been retired — see project CLAUDE.md \"V3 Performance Targets\".\n\n| Method | Measured speedup vs brute-force |\n|--------|---------------------------------|\n| Brute-force scan | Baseline |\n| HNSW (N=5,000) | ~3.2×–4.7× faster |\n| HNSW (N=20,000) | ~1.9× faster |\n| HNSW (below crossover, small N) | ties\u002Floses vs brute-force |\n| RaBitQ quantization | 32× memory reduction; 0.60 ms\u002Fquery at N≈14.7k |\n| `ruvllm_hnsw_route` (n≤11) | sub-ms per route, fixed cost |\n",{"name":5,"description":7,"allowed-tools":24,"argument-hint":25},"mcp__plugin_ruflo-core_ruflo__embeddings_generate mcp__plugin_ruflo-core_ruflo__embeddings_search mcp__plugin_ruflo-core_ruflo__embeddings_compare mcp__plugin_ruflo-core_ruflo__embeddings_init mcp__plugin_ruflo-core_ruflo__embeddings_status mcp__plugin_ruflo-core_ruflo__embeddings_hyperbolic mcp__plugin_ruflo-core_ruflo__embeddings_neural mcp__plugin_ruflo-core_ruflo__embeddings_rabitq_build mcp__plugin_ruflo-core_ruflo__embeddings_rabitq_search mcp__plugin_ruflo-core_ruflo__embeddings_rabitq_status mcp__plugin_ruflo-core_ruflo__ruvllm_hnsw_create mcp__plugin_ruflo-core_ruflo__ruvllm_hnsw_add mcp__plugin_ruflo-core_ruflo__ruvllm_hnsw_route mcp__plugin_ruflo-core_ruflo__memory_search_unified Bash","\u003Cquery> [--limit N] [--quantized]","main","d69096e239f2a89aabc2cd078c47664a9d3b1e11","04-安全渗透\u002Fruflo\u002Fplugins\u002Fruflo-agentdb\u002Fskills\u002Fvector-search",3,"https:\u002F\u002Fgithub.com\u002Flza6\u002FArmor-piercing-system-prompts","https:\u002F\u002Fgithub.com\u002Flza6","个人vx：Tf00798\r\n个人✈：@TFGY999","2026-10-08 14:29:22",{"source":8,"skillName":5,"commands":35,"prompt":45,"zipAvailable":46},[36,39,42],{"tool":37,"command":38},"npx","npx skills add lza6\u002FArmor-piercing-system-prompts --skill vector-search",{"tool":40,"command":41},"bunx","bunx skills add lza6\u002FArmor-piercing-system-prompts --skill vector-search",{"tool":43,"command":44},"pnpm","pnpm dlx skills add lza6\u002FArmor-piercing-system-prompts --skill vector-search","安装技能，技能地址：https:\u002F\u002Fgithub.com\u002Flza6\u002FArmor-piercing-system-prompts\u002Fblob\u002Fmain\u002F04-安全渗透\u002Fruflo\u002Fplugins\u002Fruflo-agentdb\u002Fskills\u002Fvector-search\u002FSKILL.md",true,[],[49,57,62,69,76,83],{"skillNo":50,"name":51,"nameZh":6,"description":52,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":53,"categories":54,"keywords":55,"publishedAt":56},"01M7FFG485KY83TD6HSC4YPCMC","obsidian-clipper-template-creator","Guide for creating templates for the Obsidian Web Clipper. Use when you want to create a new clipping template, understand available variables, or format clipped content.","https:\u002F\u002Fgithub.com\u002Flza6\u002FArmor-piercing-system-prompts\u002Fblob\u002Fmain\u002F07-后端与架构\u002Fobsidian-clipper-template-creator\u002FSKILL.md",[],[],"2026-11-16 10:05:34",{"skillNo":58,"name":51,"nameZh":6,"description":52,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":59,"categories":60,"keywords":61,"publishedAt":56},"01M7FFG0T6AZ0STEJPSB3CX84K","https:\u002F\u002Fgithub.com\u002Flza6\u002FArmor-piercing-system-prompts\u002Fblob\u002Fmain\u002F07-后端与架构\u002Fobsidian-clipper-template-creator-antigravity-awesome-skills-main\u002FSKILL.md",[],[],{"skillNo":63,"name":64,"nameZh":6,"description":65,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":66,"categories":67,"keywords":68,"publishedAt":56},"01M7FFFY0QEM14SKWPDYGMXM2P","python-api","Programmatic Object-Detection-Metrics API guidance for BoundingBox, BoundingBoxes, Evaluator, enums, coordinate conversion, AP\u002FmAP semantics, and source-style import diagnostics.","https:\u002F\u002Fgithub.com\u002Flza6\u002FArmor-piercing-system-prompts\u002Fblob\u002Fmain\u002F07-后端与架构\u002Fobject-detection-metrics\u002Fsub-skills\u002Fpython-api\u002FSKILL.md",[],[],{"skillNo":70,"name":71,"nameZh":6,"description":72,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":73,"categories":74,"keywords":75,"publishedAt":56},"01M7FFFVZQZ2GJ0C0TGP7JHD20","file-evaluation","Folder-based PASCAL VOC AP\u002FmAP evaluation from ground-truth and detection text files.","https:\u002F\u002Fgithub.com\u002Flza6\u002FArmor-piercing-system-prompts\u002Fblob\u002Fmain\u002F07-后端与架构\u002Fobject-detection-metrics\u002Fsub-skills\u002Ffile-evaluation\u002FSKILL.md",[],[],{"skillNo":77,"name":78,"nameZh":6,"description":79,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":80,"categories":81,"keywords":82,"publishedAt":56},"01M7FFFSFNYSE77SJ6PE1CK2DA","object-detection-metrics","Guides legacy Object-Detection-Metrics PASCAL VOC AP\u002FmAP evaluation from detection text folders and direct Python API classes.","https:\u002F\u002Fgithub.com\u002Flza6\u002FArmor-piercing-system-prompts\u002Fblob\u002Fmain\u002F07-后端与架构\u002Fobject-detection-metrics\u002FSKILL.md",[],[],{"skillNo":84,"name":85,"nameZh":6,"description":86,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":87,"categories":88,"keywords":89,"publishedAt":56},"01M7FFFP0GCK1RVXBCPQ8APBMT","nx-workspace-patterns","Configure and optimize Nx monorepo workspaces. Use when setting up Nx, configuring project boundaries, optimizing build caching, or implementing affected commands.","https:\u002F\u002Fgithub.com\u002Flza6\u002FArmor-piercing-system-prompts\u002Fblob\u002Fmain\u002F07-后端与架构\u002Fnx-workspace-patterns\u002FSKILL.md",[],[]]