Update to realLaw AI

{
  "name": "uae-legislation-research",
  "description": "Use when a user asks factual questions about UAE legislation, legal compliance, tax rates, thresholds, exemptions, obligations, deadlines, penalties, or needs to find, review, compare, cite, or summarize published UAE legislation, legal news, or digests through the realLaw AI MCP.",
  "included_files": [
    {
      "relative_path": "agents/openai.yaml",
      "size_in_bytes": 626
    }
  ],
  "skill_md_contents": "---\nname: uae-legislation-research\ndescription: Use when a user asks factual questions about UAE legislation, legal compliance, tax rates, thresholds, exemptions, obligations, deadlines, penalties, or needs to find, review, compare, cite, or summarize published UAE legislation, legal news, or digests through the realLaw AI MCP.\n---\n\n# UAE Legislation Research\n\nUse realLaw AI for published UAE legislation evidence and answer in the user's language.\n\n## Core rules\n\n- Write the final answer in the user's language, but send concise English search phrases to tools.\n- When realLaw AI is selected, mentioned, or this skill is invoked, retrieve evidence before answering substantive UAE legislation, compliance, or tax questions. Never answer them from model memory alone; retrieve each concept in a multi-topic question.\n- Preserve official titles, law numbers, years, and citations. Do not invent an official translation.\n- Use `language: en` for every legislation tool that accepts `language`; do not add it to `search_legislation_keywords` or `list_legislation_filters`. Translate retrieved English evidence only in the final answer.\n- Ordinary factual questions, summaries, explanations, comparisons, citations, and news or digest presentations require complete retrieval of every selected or presented document through `hasMore=false`.\n- If intent is ambiguous, treat it as content-bearing and completely retrieve every selected or presented document.\n- Complete retrieval may be skipped only when the user explicitly requests metadata-only output limited to `title`, `sourceRef`, `releaseDate`, and `sourceUrl`. Return only those fields; `createdAt` is not part of this exception.\n- Snippets and compact matches are discovery-only. Use them to identify, rank, or discard candidates, not for content or direct relevance claims.\n- Do not infer missing metadata, legal effect, or identity. If no relevant document can be completely retrieved, say the corpus could not verify the answer; do not substitute model memory or the public web unless separately requested.\n- Cite every substantive conclusion with a returned `sourceUrl`; never construct or alter one.\n\n## Route the request\n\n| User intent | First tool | Follow-up |\n|---|---|---|\n| Exact title, number, year, issue number, or citation | `find_law_document_by_reference` | Resolve, then retrieve selected matches completely |\n| Latest or topic updates, news, or digests | `get_legislation_news_digests` | Use both sections, then retrieve presented items completely |\n| Factual question without an exact reference | `search_legislation_keywords` | Search each concept, then use returned phrases in `search_legislation` |\n| Broad legal topic | `search_legislation_keywords` | Pass a matching phrase to `search_legislation` |\n| Topic with corpus filters | `list_legislation_filters` and `search_legislation_keywords` | Pass returned keys and keywords to `search_legislation` |\n| Full text, summary, citation, comparison, or presented result | `get_law_document` | Continue with `nextCursor` until `hasMore=false` |\n\n## Search topics and filters\n\n1. Extract topic, mandatory constraints, dates, and output. `search_legislation` requires `query`; ask when a filtered request has no topic.\n2. Translate to short English legal phrases and split multi-part questions into two to four concepts. Discover each with `search_legislation_keywords`; use only matching returned keywords.\n3. Resolve uncertain filters with `list_legislation_filters`, mapping `jurisdiction`, `authority`, `documentType`, and `fieldOfLaw` to their plural search arrays.\n4. Pass returned keys exactly; never guess IDs or send frontend parameters such as `zone[]` or `authority[]`.\n5. Supported `statuses`: `in_force`, `not_in_force`, `will_come_into_force`; `contentTypes`: `laws_and_regulations`. Preserve dates and separate phrases that would over-constrain one query.\n\n## Handle special routes\n\n- For news without a topic, omit digest `query`. For topic news, use concise English and preserve `fromDate` and `toDate`; without dates the tool uses the last 14 calendar dates, including today.\n- Treat `topicCandidates` as unverified discovery candidates. Completely retrieve every candidate selected for presentation before calling it a direct topic match or describing its content. Discarded candidates do not require complete retrieval.\n- Always include compact `recentDigests`; topic `query` never filters it. Completely retrieve each presented digest unless the four-field metadata-only exception applies.\n- If `noTopicCandidatesMessageRequired=true`, or no candidate remains relevant after complete inspection, state that no direct matches were found and label `recentDigests` as broader period coverage, never direct matches.\n- When `recentDigestsTruncated=true`, disclose that additional digest documents exist beyond the returned limit.\n- Use `period.dateField=releaseDate`: `releaseDate` is the issue date; `createdAt` is the edition-record creation timestamp, not the issue date.\n- An exact reference match is identity metadata, not text. Completely retrieve it before ordinary use or presentation unless the four-field metadata-only exception applies.\n- Select an ambiguous reference only when title, number, and year establish the match; otherwise clarify or offer related-topic search.\n\n## Read document text\n\n1. Call `get_law_document` with returned `documentRef`, `language: en`, and `maxCharacters: 50000`; omit `cursor` first.\n2. While `hasMore: true`, pass `nextCursor` unchanged with the same arguments.\n3. Continue until `hasMore: false`; never stop because a fragment appears sufficient, even for a specific provision.\n4. Do not use or present any document-based content claim before `hasMore: false`.\n5. If a continuation call fails or client limits prevent completion, state that the full document was not retrieved and omit all content claims based on its partial fragments.\n6. Keep cursors internal. Link only the returned `sourceUrl`; never derive a URL.\n\n## Recover from weak results\n\n1. Retry weak results with simpler legal synonyms, validated phrases, and verified filter keys. Relax at most one optional filter; never relax a mandatory constraint silently.\n2. If a citation remains unresolved, report that. Search related documents only with consent; otherwise ask for a narrower topic, citation, authority, jurisdiction, or date range.\n\n## Return the answer\n\nCompletely retrieve every selected or presented document, including ordinary lists, unless the user explicitly requests metadata-only output limited to `title`, `sourceRef`, `releaseDate`, and `sourceUrl`; discarded candidates need not be read. Report relevance, summaries, content, changes, or other metadata only after completion. Cite each conclusion with returned `sourceUrl`. Disclose incomplete retrieval and omit that document's content claims. Label ambiguities, missing fields, and broadened searches.\n"
}