菜单
Cast a broad net across the entire library — semantic search, conference talks, cross-volume connections. Surfaces patterns and relationships that deep reading misses. Use after deep-reading to widen the study, or independently for exploratory research.
用 Codex 或 Claude 帮你安装 复制这段 Prompt，粘贴到 Codex、Claude 或其他助手里，让它检查 Skill 页面并帮你完成安装。
基于 SOC 职业分类
Verify scripture and talk quotes against actual source files. Enforces read-before-quoting, the cite count rule, and the pre-publish checklist. Load when writing ANY document that cites scriptures, conference talks, transcripts, or other verifiable sources — studies, evaluations, lessons, guides, docs, everything.
Look up which General Conference talks, Journal of Discourses entries, and other sources cite a particular scripture verse using the BYU Scripture Citation Index (scriptures.byu.edu). Use when you want to know who has taught about a verse, how frequently a verse appears in conference, or to discover that a verse has surprisingly little modern commentary.
Stress-test the study's arguments, assumptions, and gaps before finalizing. Asks 'what are we missing?', 'where are we overreaching?', and 'does the text actually say what we think it says?' Use between gap analysis and first draft to sharpen the study.
Telescope into a specific scripture passage or narrow topic. Follow every footnote, trace every cross-reference, read full chapters for context. Produces intermediate findings in a working document. Use for focused, depth-first study on a single text or concept.
Six text-checkable properties for evaluating non-canonical sources (books, articles, transcripts, web content) before citing or recommending them. Loaded passively whenever the agent is asked to assess, summarize, recommend, or quote from a source outside the LDS canon. Names what the agent CAN check (correspondence, coherence, canonical fit, pillars-rhetoric, mark-keeping, steelman) and what only the user with the Spirit can check (witness, soul-fruit, animating spirit, personal stewardship-fit, override). Grounded in study/discernment-and-the-comprehending-eye.md.
Run the publish script to convert study documents to public HTML, then git add, commit, and push. Use when the user says publish, commit, push, or asks to share their work.
| name | wide-search |
| description | Cast a broad net across the entire library — semantic search, conference talks, cross-volume connections. Surfaces patterns and relationships that deep reading misses. Use after deep-reading to widen the study, or independently for exploratory research. |
| argument-hint | [concept or theme to search broadly] |
Wide search is the wide-angle lens — scanning across all five standard works, 54 years of conference talks, manuals, and existing studies for connections, patterns, and unexpected resonances. This skill produces the kind of discovery where you search "lifted up" and find it connects Numbers 21, John 3, Helaman 8, 3 Nephi 27, AND a 1978 conference talk you never would have found manually.
This contrasts with the deep-reading skill, which telescopes into one passage. Both are valuable. Wide search usually follows deep reading in a phased study.
If this follows a deep-reading phase, start with the "Threads to Pull" from the intermediate findings. These are your search seeds.
If this is standalone, articulate 3-5 search queries before starting:
Use mcp__gospel-engine-v2__gospel_search with mode: "keyword" (the default) for specific phrases, names, or terms:
Remember: These results are pointers. Note file paths. Do NOT quote from search results.
Use mcp__gospel-engine-v2__gospel_search with mode: "semantic" for conceptual connections:
Semantic search finds what keyword search misses. "Serpent as a type of Christ" won't match keyword search, but semantic search finds Numbers 21, John 3, and Helaman 8 together.
For the best of both, use mode: "hybrid" — keyword candidates reranked by semantic similarity. Use source: "conference" to scope to talks; filter by speaker, year_from, year_to, or TITSW dimensions when relevant.
Search study/*.md (with Grep) for related work. The study corpus is an interconnected web — previous studies may have already traced threads relevant to this topic.
For every promising discovery, Read the actual source. The source-verification skill applies here — search results are pointers, not sources.
Add findings to the working document (if phased study) or compile for synthesis:
## Wide Search Findings
### Cross-Volume Pattern: [pattern name]
- OT: [reference + insight]
- NT: [reference + insight]
- BoM: [reference + insight]
- D&C: [reference + insight]
- PoGP: [reference + insight]
### Conference Talks
- [Speaker, Title, Date](link) — key quote and connection
- [Speaker, Title, Date](link) — key quote and connection
### Unexpected Connections
- [Something you didn't expect to find]
### Connections to Past Studies
- [study/topic.md](link) — how this connects
| Regular Discovery (One-Shot) | Wide Search Phase |
|---|---|
| Searches to find what to read | Searches to find connections |
| Stops when enough sources found | Keeps going to find the web |
| One round of search | Multiple rounds with different query strategies |
| Keyword OR semantic | Both keyword AND semantic, deliberately |
| Results go straight to study | Results go to working notes, then to synthesis |
gospel_search is one tool with three modes. Each finds different things:
mode: "keyword" | mode: "semantic" | mode: "hybrid" |
|---|---|---|
| Finds exact phrases | Finds related concepts | Keyword candidates, semantic rerank |
| Good for: names, specific terms, quoted phrases | Good for: themes, parallels, "verses about X" | Good for: best precision + recall |
| Misses: paraphrases, different vocabulary | Misses: exact matches, specific rare terms | More expensive — use when you want both |
A study that only uses one mode misses entire classes of results. Run keyword first, then semantic on the same concept; if results diverge wildly, run combined.
In a phased study plan:
Wide search feeds synthesis with the breadth and cross-references that make studies like serpent-and-dragon.md comprehensive — 24 source files across all 5 standard works.