[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"skill:01M7ADAH24ANDE7H4CJ2KWFPF1":3,"skill-same:01M7ADAH24ANDE7H4CJ2KWFPF1":49},{"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":27,"commitSha":28,"skillDirectory":29,"forksCount":30,"repositoryHtmlUrl":31,"repositoryHomepageUrl":32,"ownerHtmlUrl":33,"ownerBlogUrl":6,"ownerDescription":6,"updatedAt":34,"install":35,"resources":48},"01M7ADAH24ANDE7H4CJ2KWFPF1","agentic-rag",null,"Use when implementing an agentic RAG product in a target project (support desk, knowledge assistant, evidence-gated Q&A). Playbook for the retrieve -> decision-gate -> answer-or-escalate architecture with multimodal index and human-review loop. Builds the system; does not answer questions with it.","Leostruka\u002Fdevin-bundle","Leostruka","https:\u002F\u002Favatars.githubusercontent.com\u002Fu\u002F81370685?v=4","01M7AD8MAK63VDNAWRDRRRW04N","leandro ostruka","01M7AD8MBS65YDS1KMMV7YPM07",0,"MIT","https:\u002F\u002Fgithub.com\u002FLeostruka\u002Fdevin-bundle\u002Fblob\u002Fmain\u002Fskills\u002Fagentic-rag\u002FSKILL.md",[],[],"2026-11-15 14:19:17",false,"---\nname: agentic-rag\ndescription: Use when implementing an agentic RAG product in a target project (support desk, knowledge assistant, evidence-gated Q&A). Playbook for the retrieve -> decision-gate -> answer-or-escalate architecture with multimodal index and human-review loop. Builds the system; does not answer questions with it.\ntriggers: [user, model]\n---\n\n# Agentic RAG (evidence-gated Q&A)\n\n## Overview\n\nImplementation playbook for the agentic RAG support-desk architecture\n(source: Daily Dose of DS \"Jev-powered Agentic RAG\" walkthrough). The\nsystem retrieves evidence, asks a decision model whether the evidence is\nsufficient, answers with citations when it is, and escalates to a human\nticket when it is not. Human replies feed back into the index.\n\nFive slots, stack-agnostic:\n\n| Slot | Reference choice | Swappable with |\n|---|---|---|\n| Backend | InsForge OSS (Postgres+pgvector, auth, storage, edge functions, realtime, model gateway) | Supabase \u002F plain Postgres+pgvector + functions |\n| Embeddings | EmbeddingGemma 2 (multimodal, 768-dim, text+image same space) | Any text embedder if corpus is text-only |\n| Decision model | Laya multilingual (Convai Innovations; 322M BERT-style encoder, Apache 2.0, self-host, free; this bundle already ships `laya-tools`\u002F`layad` and the `implement-laya` skill) | Free\u002Fopen: Kev 9B, Jeeves 9B, Strands Decider 2B, open-weight d1-3B\u002Fd1-omni-600M. Paid optional: Jev or hosted d1\u002Fd1-ultra (Choice\u002FScore\u002FNoul API, text+image). Last resort: LLM-judge with strict JSON schema |\n| Generator | Vision-capable LLM through a model gateway, model id in secrets | Any cited-generation capable model |\n| Hosting | One L4 (~7.2 GB both models); SIE for multi-model single-GPU serving | Any GPU host or managed endpoint |\n\n## Architecture selector (when to use what)\n\n| Corpus & workload | Pattern | Why |\n|---|---|---|\n| Static corpus \u003C200k tokens (~500 pages), low churn | Prompt context + prompt caching (no RAG) | Zero retrieval infra or chunk-boundary loss |\n| Single-source docs + per-user chat history | Linear RAG + cross-encoder reranker | One vector search (top-20) + reranker (`max_results = 5`); predictable latency |\n| Long structured docs (filings, contracts, manuals) with real ToC | Vectorless \u002F tree-index retrieval (PageIndex-style), optionally feeding this playbook's gate | Similarity fails on template-identical sections and cross-references; tree lets the model follow them like an analyst |\n| High-stakes Q&A, multi-source (docs\u002FSQL\u002FAPI), support desk | **Agentic RAG (this playbook)** | Decision gate blocks ungrounded generation, allows 1 wider retry, and escalates to human with feedback loop |\n\n### Vectorless retrieval notes (source: Daily Dose of DS follow-up)\n\n- Failure modes it fixes: chunks from different periods\u002Fversions that\n  look identical to an embedder (wrong-year answers, silent), and\n  pointers like \"see Appendix G\" whose target never ranks.\n- Trade: each tree hop is an LLM call (latency\u002Fcost per query up),\n  index build costs minutes of LLM calls per document, and it only\n  works when documents have real structure. Lots of short unrelated\n  text stays on the vector side; period-mixup can also be mitigated\n  with metadata filters there.\n- If adopted here, the tree replaces only the retrieval stage --\n  decision gate, escalation, and human loop stay identical. Version\n  the tree with the document it was built from: a stale summary node\n  is a silent wrong-version answer.\n\n## Context separation (Conversation vs. Memory vs. RAG)\n\nKeep the three context sources distinct in storage metadata and in the generator system prompt:\n\n| Pattern | What it stores | Metadata \u002F scope | Prompt rule |\n|---|---|---|---|\n| Conversation storage | Last $N$ verbatim turns (e.g. 6) | Passed in `Context` array | Resolves follow-ups (\"how do I do that in Docker?\") that lack semantic overlap |\n| Long-term user memory | Earlier turns or extracted user facts | `kind = conversation`, `user_id` from auth principal | Attribute to the user only when `kind = conversation` |\n| External knowledge (RAG) | Shared corpus chunks + page PNGs | `kind = document`, `title`, `url`, `page` | Technical facts only; never attribute document examples to the caller |\n\n## Iron rules\n\n1. Never call the generation model when the gate fails. Escalation is a\n   first-class output, not an error.\n2. Thresholds live in a DB\u002Fconfig table, separate for text and image\n   candidates; tuning must never require a deploy.\n3. Log every gate decision (candidates, scores, verdict) to\n   `decision_log` before answering or escalating.\n4. Index each page twice: ~220-word text chunks AND a rendered page PNG,\n   both tagged with the page number. Merge candidates by page.\n5. Answers must cite page numbers and return thumbnails\u002Fexcerpts when the\n   channel supports it. The generator answers only from supplied evidence.\n6. One wider retry max before ticketing: widen top-k\u002Ffilters once, never\n   loop.\n7. Human actions stay human: reps write replies, customer receives email.\n   \"Save to knowledge\" re-indexes the Q&A pair as a new chunk.\n8. Tenant and user isolation via RLS \u002F authenticated principal (`ClaimsPrincipal` \u002F JWT):\n   never accept `user_id` or `session_id` from the request body, and escape or parameterize all metadata filters.\n9. No silent rerank or gate fallback: inspect stream status events (`RERANKING_FAILED`, `SUMMARIZATION_FAILED`) and verify the `rerank`\u002F`gate` stage actually ran; enforce `max_results` cut even when reranking is off.\n10. End-to-end stream deadline and reasoning token budget: `HttpClient.Timeout` only covers response headers on streamed replies; link a request-wide deadline across history load + stream read + turn store, and size `gen_token_budget` (e.g. >=2048) so reasoning models do not exhaust budget on thinking tokens.\n\n## Phase route\n\n| Gate | Name | Exit condition |\n|---|---|---|\n| P0 | Intake | Corpus shape (docs, pages, languages), tenancy model, answer channels, threshold starting values, human-review SLA defined in `.devin` notes or issue. |\n| G1 | Data model | Tables for documents, chunks (vector dim per embedder), conversations, tickets, decision_log, thresholds; tenant-scoped RLS verified with two-tenant test. |\n| G2 | Index pipeline | Upload -> storage -> job -> per-page text chunks + page PNG -> embed both -> persist with page tags; deterministic doc IDs + async ingestion status polling (`COMPLETED`\u002F`FAILED`); 20-page PDF indexes in \u003C1 min. |\n| G3 | Query pipeline + gate | Embed question (+attached image) -> dual search (text, image) -> merge by page -> decision model rerank+answerability -> thresholds -> answer w\u002F citations OR one wider retry OR ticket. Gate unit-tested above\u002Fbelow threshold and on stage failure. |\n| G4 | Human loop | Ticket appears realtime with question + rejected candidates + scores; rep reply emails customer; save-to-knowledge re-indexes. |\n| G5 | Eval | Held-out Q&A set: retrieval ranking (`hit@1`, `hit@3`, `MRR`), gate precision\u002Frecall, citation correctness, escalation rate, and split `p50`\u002F`p95` latency (retrieval vs. rerank\u002Fgate vs. LLM generation) reported; thresholds tuned from data. |\n\n## Acceptance criteria\n\n- Query with sufficient evidence: answer + page citations + decision_log row.\n- Query with insufficient evidence: single wider retry, then ticket with\n  candidate scores visible to the rep; generation model never called.\n- Rep reply reaches customer by email; save-to-knowledge makes the same\n  question retrievable.\n- Threshold change in DB alters gate behavior without a deploy.\n\n## Non-goals\n\n- No auto-submit\u002Fauto-send to end users without rep confirmation.\n- No figure-level extraction (full-page PNG only; vector diagrams lose\n  captions otherwise).\n- No unbounded retry loops, no answer on sub-threshold evidence.\n\n## Integration notes\n\n- `data-analyst`\u002F`research` skills can reuse the gate pattern for\n  evidence sufficiency before answering.\n- On this bundle host, `extensions\u002Fvoicebox` (tts) + `asr` can add voice\n  to the desk frontend; optional, not part of the core pipeline.\n","# Agentic RAG (evidence-gated Q&A)\n\n## Overview\n\nImplementation playbook for the agentic RAG support-desk architecture\n(source: Daily Dose of DS \"Jev-powered Agentic RAG\" walkthrough). The\nsystem retrieves evidence, asks a decision model whether the evidence is\nsufficient, answers with citations when it is, and escalates to a human\nticket when it is not. Human replies feed back into the index.\n\nFive slots, stack-agnostic:\n\n| Slot | Reference choice | Swappable with |\n|---|---|---|\n| Backend | InsForge OSS (Postgres+pgvector, auth, storage, edge functions, realtime, model gateway) | Supabase \u002F plain Postgres+pgvector + functions |\n| Embeddings | EmbeddingGemma 2 (multimodal, 768-dim, text+image same space) | Any text embedder if corpus is text-only |\n| Decision model | Laya multilingual (Convai Innovations; 322M BERT-style encoder, Apache 2.0, self-host, free; this bundle already ships `laya-tools`\u002F`layad` and the `implement-laya` skill) | Free\u002Fopen: Kev 9B, Jeeves 9B, Strands Decider 2B, open-weight d1-3B\u002Fd1-omni-600M. Paid optional: Jev or hosted d1\u002Fd1-ultra (Choice\u002FScore\u002FNoul API, text+image). Last resort: LLM-judge with strict JSON schema |\n| Generator | Vision-capable LLM through a model gateway, model id in secrets | Any cited-generation capable model |\n| Hosting | One L4 (~7.2 GB both models); SIE for multi-model single-GPU serving | Any GPU host or managed endpoint |\n\n## Architecture selector (when to use what)\n\n| Corpus & workload | Pattern | Why |\n|---|---|---|\n| Static corpus \u003C200k tokens (~500 pages), low churn | Prompt context + prompt caching (no RAG) | Zero retrieval infra or chunk-boundary loss |\n| Single-source docs + per-user chat history | Linear RAG + cross-encoder reranker | One vector search (top-20) + reranker (`max_results = 5`); predictable latency |\n| Long structured docs (filings, contracts, manuals) with real ToC | Vectorless \u002F tree-index retrieval (PageIndex-style), optionally feeding this playbook's gate | Similarity fails on template-identical sections and cross-references; tree lets the model follow them like an analyst |\n| High-stakes Q&A, multi-source (docs\u002FSQL\u002FAPI), support desk | **Agentic RAG (this playbook)** | Decision gate blocks ungrounded generation, allows 1 wider retry, and escalates to human with feedback loop |\n\n### Vectorless retrieval notes (source: Daily Dose of DS follow-up)\n\n- Failure modes it fixes: chunks from different periods\u002Fversions that\n  look identical to an embedder (wrong-year answers, silent), and\n  pointers like \"see Appendix G\" whose target never ranks.\n- Trade: each tree hop is an LLM call (latency\u002Fcost per query up),\n  index build costs minutes of LLM calls per document, and it only\n  works when documents have real structure. Lots of short unrelated\n  text stays on the vector side; period-mixup can also be mitigated\n  with metadata filters there.\n- If adopted here, the tree replaces only the retrieval stage --\n  decision gate, escalation, and human loop stay identical. Version\n  the tree with the document it was built from: a stale summary node\n  is a silent wrong-version answer.\n\n## Context separation (Conversation vs. Memory vs. RAG)\n\nKeep the three context sources distinct in storage metadata and in the generator system prompt:\n\n| Pattern | What it stores | Metadata \u002F scope | Prompt rule |\n|---|---|---|---|\n| Conversation storage | Last $N$ verbatim turns (e.g. 6) | Passed in `Context` array | Resolves follow-ups (\"how do I do that in Docker?\") that lack semantic overlap |\n| Long-term user memory | Earlier turns or extracted user facts | `kind = conversation`, `user_id` from auth principal | Attribute to the user only when `kind = conversation` |\n| External knowledge (RAG) | Shared corpus chunks + page PNGs | `kind = document`, `title`, `url`, `page` | Technical facts only; never attribute document examples to the caller |\n\n## Iron rules\n\n1. Never call the generation model when the gate fails. Escalation is a\n   first-class output, not an error.\n2. Thresholds live in a DB\u002Fconfig table, separate for text and image\n   candidates; tuning must never require a deploy.\n3. Log every gate decision (candidates, scores, verdict) to\n   `decision_log` before answering or escalating.\n4. Index each page twice: ~220-word text chunks AND a rendered page PNG,\n   both tagged with the page number. Merge candidates by page.\n5. Answers must cite page numbers and return thumbnails\u002Fexcerpts when the\n   channel supports it. The generator answers only from supplied evidence.\n6. One wider retry max before ticketing: widen top-k\u002Ffilters once, never\n   loop.\n7. Human actions stay human: reps write replies, customer receives email.\n   \"Save to knowledge\" re-indexes the Q&A pair as a new chunk.\n8. Tenant and user isolation via RLS \u002F authenticated principal (`ClaimsPrincipal` \u002F JWT):\n   never accept `user_id` or `session_id` from the request body, and escape or parameterize all metadata filters.\n9. No silent rerank or gate fallback: inspect stream status events (`RERANKING_FAILED`, `SUMMARIZATION_FAILED`) and verify the `rerank`\u002F`gate` stage actually ran; enforce `max_results` cut even when reranking is off.\n10. End-to-end stream deadline and reasoning token budget: `HttpClient.Timeout` only covers response headers on streamed replies; link a request-wide deadline across history load + stream read + turn store, and size `gen_token_budget` (e.g. >=2048) so reasoning models do not exhaust budget on thinking tokens.\n\n## Phase route\n\n| Gate | Name | Exit condition |\n|---|---|---|\n| P0 | Intake | Corpus shape (docs, pages, languages), tenancy model, answer channels, threshold starting values, human-review SLA defined in `.devin` notes or issue. |\n| G1 | Data model | Tables for documents, chunks (vector dim per embedder), conversations, tickets, decision_log, thresholds; tenant-scoped RLS verified with two-tenant test. |\n| G2 | Index pipeline | Upload -> storage -> job -> per-page text chunks + page PNG -> embed both -> persist with page tags; deterministic doc IDs + async ingestion status polling (`COMPLETED`\u002F`FAILED`); 20-page PDF indexes in \u003C1 min. |\n| G3 | Query pipeline + gate | Embed question (+attached image) -> dual search (text, image) -> merge by page -> decision model rerank+answerability -> thresholds -> answer w\u002F citations OR one wider retry OR ticket. Gate unit-tested above\u002Fbelow threshold and on stage failure. |\n| G4 | Human loop | Ticket appears realtime with question + rejected candidates + scores; rep reply emails customer; save-to-knowledge re-indexes. |\n| G5 | Eval | Held-out Q&A set: retrieval ranking (`hit@1`, `hit@3`, `MRR`), gate precision\u002Frecall, citation correctness, escalation rate, and split `p50`\u002F`p95` latency (retrieval vs. rerank\u002Fgate vs. LLM generation) reported; thresholds tuned from data. |\n\n## Acceptance criteria\n\n- Query with sufficient evidence: answer + page citations + decision_log row.\n- Query with insufficient evidence: single wider retry, then ticket with\n  candidate scores visible to the rep; generation model never called.\n- Rep reply reaches customer by email; save-to-knowledge makes the same\n  question retrievable.\n- Threshold change in DB alters gate behavior without a deploy.\n\n## Non-goals\n\n- No auto-submit\u002Fauto-send to end users without rep confirmation.\n- No figure-level extraction (full-page PNG only; vector diagrams lose\n  captions otherwise).\n- No unbounded retry loops, no answer on sub-threshold evidence.\n\n## Integration notes\n\n- `data-analyst`\u002F`research` skills can reuse the gate pattern for\n  evidence sufficiency before answering.\n- On this bundle host, `extensions\u002Fvoicebox` (tts) + `asr` can add voice\n  to the desk frontend; optional, not part of the core pipeline.\n",{"name":5,"triggers":24,"description":7},[25,26],"user","model","main","02b95ddb62959a00b3c31442c2046aca0d19563c","skills\u002Fagentic-rag",1,"https:\u002F\u002Fgithub.com\u002FLeostruka\u002Fdevin-bundle","https:\u002F\u002Fleostruka.github.io\u002Fdevin-bundle\u002F","https:\u002F\u002Fgithub.com\u002FLeostruka","2026-11-15 05:19:58",{"source":8,"skillName":5,"commands":36,"prompt":46,"zipAvailable":47},[37,40,43],{"tool":38,"command":39},"npx","npx skills add Leostruka\u002Fdevin-bundle --skill agentic-rag",{"tool":41,"command":42},"bunx","bunx skills add Leostruka\u002Fdevin-bundle --skill agentic-rag",{"tool":44,"command":45},"pnpm","pnpm dlx skills add Leostruka\u002Fdevin-bundle --skill agentic-rag","安装技能，技能地址：https:\u002F\u002Fgithub.com\u002FLeostruka\u002Fdevin-bundle\u002Fblob\u002Fmain\u002Fskills\u002Fagentic-rag\u002FSKILL.md",true,[],[50,57,64,71,78,85],{"skillNo":51,"name":52,"nameZh":6,"description":53,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":54,"categories":55,"keywords":56,"publishedAt":19},"01M7E0K3MG2XC4FSRV0MK6SFR1","montador-de-ideias","Use when assembling, expanding, or elevating a raw project idea or initial prompt into an ultra-detailed, first-principles dossier and compiled master prompt with journalistic OSINT market research, Tech Lead and Staff Architect justifications, atomic and modular hardening\u002Foptimization, and high-impact contact surfaces.","https:\u002F\u002Fgithub.com\u002FLeostruka\u002Fdevin-bundle\u002Fblob\u002Fmain\u002Fskills\u002Fmontador-de-ideias\u002FSKILL.md",[],[],{"skillNo":58,"name":59,"nameZh":6,"description":60,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":61,"categories":62,"keywords":63,"publishedAt":19},"01M7E0J7JJWZ8QMK01RZ3B3C9Z","git-github-adaptive-engineering","Use when any task touches git history, GitHub issues\u002FPRs\u002Freleases\u002FCI, repo governance, or branch and merge decisions. Runs the Fase Zero situational-awareness battery, picks one of five delivery flows via the adaptive decision engine, and enforces commit\u002FPR information-density standards. Orchestrates the gh, git-workflows, github-actions, pr, and debug\u002Ffinish skills.","https:\u002F\u002Fgithub.com\u002FLeostruka\u002Fdevin-bundle\u002Fblob\u002Fmain\u002Fskills\u002Fgit-github-adaptive-engineering\u002FSKILL.md",[],[],{"skillNo":65,"name":66,"nameZh":6,"description":67,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":68,"categories":69,"keywords":70,"publishedAt":19},"01M7E0GJFRPFEVQQMXEM58EWDM","adaptive-agenda-calendar-routine-engine","Use when planning or executing work across hours, days, weeks, or sprints. Turns the calendar into a strict time-blocked agenda (HH:MM-HH:MM blocks carrying a routine profile, expected inputs, and a handoff output artifact), runs the 5-stage cognitive flow (Radar and Hunt -> Triage plus Steve Jobs conception -> workbench manifest -> >=32-cycle loop -> Close plus Open-Minded Scrum), and unfolds macro calendar milestones into micro agenda blocks through extensions\u002Fgcal\u002Fwrapper.py.","https:\u002F\u002Fgithub.com\u002FLeostruka\u002Fdevin-bundle\u002Fblob\u002Fmain\u002Fskills\u002Fadaptive-agenda-calendar-routine-engine\u002FSKILL.md",[],[],{"skillNo":72,"name":73,"nameZh":6,"description":74,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":75,"categories":76,"keywords":77,"publishedAt":19},"01M7AE2K9FTY0BVS48N0A9JFKF","demo","Use when demonstrating a skill in a fixture project.","https:\u002F\u002Fgithub.com\u002FLeostruka\u002Fdevin-bundle\u002Fblob\u002Fmain\u002Ftests\u002Ffixtures\u002Fdevin-config\u002Fproject\u002F.devin\u002Fskills\u002Fdemo\u002FSKILL.md",[],[],{"skillNo":79,"name":80,"nameZh":6,"description":81,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":82,"categories":83,"keywords":84,"publishedAt":19},"01M7AE2G3DP38PTV3B1N7MW5TS","duplicate-skill","Use when testing duplicate skill name detection.","https:\u002F\u002Fgithub.com\u002FLeostruka\u002Fdevin-bundle\u002Fblob\u002Fmain\u002Ftests\u002Ffixtures\u002Fdevin-config\u002Fproject\u002F.devin\u002Fskills\u002Fbeta\u002FSKILL.md",[],[],{"skillNo":86,"name":80,"nameZh":6,"description":81,"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":19},"01M7AE2DBR8ZHT6T875MJ392GK","https:\u002F\u002Fgithub.com\u002FLeostruka\u002Fdevin-bundle\u002Fblob\u002Fmain\u002Ftests\u002Ffixtures\u002Fdevin-config\u002Fproject\u002F.devin\u002Fskills\u002Falpha\u002FSKILL.md",[],[]]