[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"skill:01M74MGVN4W2PYC5ZCJ266SDGQ":3,"skill-same:01M74MGVN4W2PYC5ZCJ266SDGQ":95},{"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":24,"commitSha":25,"skillDirectory":26,"forksCount":27,"repositoryHtmlUrl":28,"repositoryHomepageUrl":29,"ownerHtmlUrl":30,"ownerBlogUrl":6,"ownerDescription":31,"updatedAt":32,"install":33,"resources":46},"01M74MGVN4W2PYC5ZCJ266SDGQ","paper-search",null,"Search papers across arXiv, DBLP, OpenAlex, OpenReview, Semantic Scholar, and Crossref for a given query and year range, using .\u002Fscripts\u002Fsearch_papers.py. Use when the user asks to find papers, related work, prior art, or recent publications on a specific topic, especially when they mention a date range or specific venues like NeurIPS, ICLR, or ICML.","pengqianhan\u002FAI-Human-Research-OS","pengqianhan","https:\u002F\u002Favatars.githubusercontent.com\u002Fu\u002F4286531?v=4","01M74M3DT9HNZ6HPAK8TX69EX9","PengqianHan","01M74M3DTVSN2GXW0JH2W396FK",3,"MIT","https:\u002F\u002Fgithub.com\u002Fpengqianhan\u002FAI-Human-Research-OS\u002Fblob\u002Fmain\u002Fresearch-skills-hub\u002Fcollected-skills\u002FResearchStudio-Idea\u002Fskills\u002Fpaper_search\u002FSKILL.md",[],[],"2026-11-11 15:11:14",false,"---\nname: paper-search\ndescription: Search papers across arXiv, DBLP, OpenAlex, OpenReview, Semantic Scholar, and Crossref for a given query and year range, using .\u002Fscripts\u002Fsearch_papers.py. Use when the user asks to find papers, related work, prior art, or recent publications on a specific topic, especially when they mention a date range or specific venues like NeurIPS, ICLR, or ICML.\n---\n\n# Paper Search Skill\n\nUnified paper search across **arXiv**, **DBLP**, **OpenAlex**, **OpenReview**\n(NeurIPS \u002F ICLR \u002F ICML), **Semantic Scholar**, and **Crossref** using\n`.\u002Fscripts\u002Fsearch_papers.py`. All sources are searched **concurrently** (in\nparallel threads) by default for maximum speed. Returns results grouped by\nsource.\n\n\n## When to use\n\nTrigger this skill when the user asks things like:\n- \"Find papers on X published between 2023 and 2025.\"\n- \"Search NeurIPS \u002F ICLR \u002F ICML for work on X.\"\n- \"Get arXiv + Semantic Scholar results for X.\"\n- \"Show me recent prior art on X.\"\n\n## Inputs (all auto-inferred — NEVER ask the user for confirmation or clarification)\n\nDerive these automatically from the user's message. Run the search immediately without asking for confirmation:\n- **query**: Rephrase the user's question into a focused search phrase.\n- **start_year** (int): If the user gives a year, use it directly. If they say\n  \"last 2 years\", compute from today. Default: 2 years ago.\n- **end_year** (int): Default: current year.\n- **max_papers** (int): Number of results per source. Default: 10.\n- **sources**: Which sources to query. Default: all 6 API sources plus the\n  model-knowledge source, in this canonical order (highest-signal first, so\n  the best results render before the user scrolls):\n  `semantic_scholar open_alex arxiv openreview crossref dblp model_knowledge`.\n  Only restrict sources if the user explicitly asks.\n\n## How to run\n\nPreferred: call the CLI directly. The script lives at\n`${CLAUDE_PROJECT_DIR}\u002Fskills\u002Fpaper_search\u002Fscripts\u002Fsearch_papers.py` — invoke\nit by absolute path so the command works regardless of the current working\ndirectory (relying on `cd scripts && ...` breaks when the model is running\nfrom a different folder, which happens often).\n\nFor brevity in the examples below, treat `$SEARCH` as shorthand for that\nabsolute path:\n\n```bash\nSEARCH=\"${CLAUDE_PROJECT_DIR}\u002Fskills\u002Fpaper_search\u002Fscripts\u002Fsearch_papers.py\"\n```\n\nBasic search:\n\n```bash\npython \"$SEARCH\" \\\n    --query \"\u003CQUERY>\" \\\n    --start-year \u003CYYYY> \\\n    --end-year \u003CYYYY> \\\n    --max-papers 10\n```\n\nTo restrict to specific sources:\n\n```bash\npython \"$SEARCH\" \\\n    --query \"\u003CQUERY>\" \\\n    --start-year 2024 --end-year 2026 \\\n    --sources arxiv semantic_scholar openreview\n```\n\nTo disable parallel execution (rarely needed):\n\n```bash\npython \"$SEARCH\" \\\n    --query \"\u003CQUERY>\" \\\n    --start-year 2024 --end-year 2026 \\\n    --no-parallel\n```\n\nOr call the function directly when more control is needed (e.g. consuming the\nstructured dict rather than CLI text output). This is rarely necessary — see\n`references\u002Fprogrammatic_api.md` for the snippet.\n\n## Valid source names\n\n| Source | Key |\n|:---|:---|\n| arXiv | `arxiv` |\n| DBLP | `dblp` |\n| OpenAlex | `open_alex` |\n| OpenReview | `openreview` |\n| Semantic Scholar | `semantic_scholar` |\n| Crossref | `crossref` |\n| Model knowledge (LLM recall, no API call) | `model_knowledge` |\n\n## Output schema\n\n`search_papers()` returns a dict mapping source name to a list of paper dicts:\n\n```\n{\n  \"arxiv\": [\n    {\n      \"title\": str,\n      \"authors\": [str, ...],\n      \"year\": int,\n      \"abstract\": str,\n      \"url\": str,\n      \"venue\": str,\n      \"citation_count\": int,\n      \"publication_date\": str,\n      \"source\": str\n    }, ...\n  ],\n  \"semantic_scholar\": [...],\n  ...\n}\n```\n\nThe CLI prints results grouped by source with paper count summaries.\n\n## Output to the user\n\nAfter running, display every paper from every source, in this order:\nSemantic Scholar, OpenAlex, arXiv, OpenReview, Crossref, DBLP, Model Knowledge.\nThen provide a summary.\n\nWhy full recall matters: users invoking this skill are doing literature reviews,\nrelated-work surveys, or prior-art checks. The value comes from seeing the\ncomplete set of hits — a missed paper can mean a missed citation or a\nduplicated research effort. Summaries are meant to *augment* the full tables,\nnot replace them, so don't collapse results into a digest \"to save space.\"\nThe user can skim; they can't un-skip a paper they never saw.\n\n### Step 1: Display ALL results from every source\n\nFor **each source**, display every paper in a **markdown table** with title, date, venue, and citation count.\n\nFormat each source as a table grouped under a source heading, e.g.:\n\n```\n### arXiv (N papers)\n\n| #   | Title       | Date    | Venue   | Citations |\n|-----|-------------|---------|---------|-----------|\n| [1](paper url) | Title here | 2024-03 | NeurIPS | 42 |\n| [2](paper url) | Title here | 2023-11 | ICLR    | 10 |\n```\n\nIf a source returned 0 results, note it explicitly\n(e.g. \"### OpenReview (0 papers) — No matches found in this window\").\n\nIf errors occurred during search, they are printed to stderr by the script —\nsurface them to the user, never hide them.\n\n### Step 2: Summary of all searched results\n\nAfter displaying all papers, provide a **comprehensive summary** with the\nfollowing sections, in this exact order:\n\n1. **Overview**: query used, year range, and total number of papers found. One or\n   two sentences framing what the corpus covers.\n2. **Trends**: Temporal patterns (e.g. \"interest surged in 2024\"), dominant\n   venues, methodological shifts, and recurring author groups or labs.\n3. **Key themes**: 3–6 main research themes \u002F clusters across all results,\n   each with a one-line description and 2–3 representative paper numbers.\n4. **Keywords frequency**: A table of the most frequent technical terms \u002F\n   concepts extracted from titles and abstracts, with counts. Format:\n   `| Keyword | Count |`. Include the top 5.\n5. **Most cited by accepted paper**: Top 5 most-cited accepted papers across all sources,\n   ranked by citation count, as a table: `| Rank | Title | Year | Citations |`.\n6. **Most cited by first author**: Top 5 first authors ranked by total citations\n   accumulated across papers in this result set, as a table:\n   `| Rank | Author | Papers in set | Total citations |`.\n   The **Author** column must contain ONLY the author's name (e.g. `Jane Doe`).\n   Do not append paper titles, affiliations, venues, or any other information\n   in this column — paper counts and citation totals live in their own columns.\n7. **Recommendations for reading**: 3–5 papers most relevant and impactful to the user's\n   original query, ordered as a reading path (foundational → recent), each\n   with a one-line justification.\n\n\n## Dependencies & failure modes\n\n- **arXiv**: stdlib only (uses arXiv API).\n- **DBLP**: uses DBLP API.\n- **OpenAlex**: uses OpenAlex API.\n- **OpenReview**: requires `pip install openreview-py`.\n- **Semantic Scholar**: uses Semantic Scholar API.\n- **Crossref**: uses Crossref API.\n- **Model knowledge**: no API call. Papers are recalled from the model's own\n  training data — fast and free, but capped by the model's knowledge cutoff\n  and prone to hallucination. See the \"Model knowledge source\" section below\n  for how to use it responsibly.\n\nIf a source fails, `search_papers` catches the exception, prints the error, and\ncontinues with the remaining sources. Never retry blindly; report errors to the\nuser.\n\n## Model knowledge source\n\nThe `model_knowledge` source is different from the others: it has no API and\nno script call. Instead, after the CLI search returns, recall 5–10 additional\npapers from your own training data that match the query and year range, and\npresent them as a separate source in the output.\n\n### Why include it\n\nAPI search is high-precision but low-recall in two predictable cases:\n1. **Foundational older papers** that practitioners always cite but that\n   keyword search misses (e.g. the original BERT or ResNet paper when the\n   query is about a recent variant).\n2. **Cross-disciplinary classics** that live in venues the APIs index poorly.\n\nModel recall complements the APIs by surfacing the \"everyone knows this one\"\npapers that don't always come back from a fresh keyword query.\n\n### How to populate it\n\nAfter the CLI run completes:\n\n1. Reflect on what you know about the query topic.\n2. List up to 10 papers from your training data that fit the query and year\n   range, with: title, primary author(s), year, venue, and a one-line reason\n   it's relevant.\n3. Deduplicate against the API results — if a paper already appeared in any\n   API source, do not repeat it under `model_knowledge`.\n4. Flag confidence honestly. The model knowledge column has no citation count\n   and no live URL; if you're not sure a paper exists exactly as you remember\n   it, mark it `(uncertain — verify)` in the table rather than presenting it\n   as confirmed.\n\n### Why honesty matters here\n\nHallucinated paper titles are the classic LLM failure mode for this task. A\nfake \"Smith et al., 2023, NeurIPS\" looks identical to a real one in a\nmarkdown table, and the user has no way to tell. The point of this source is\nto surface *real* papers the APIs missed — not to pad the list. If you can't\nrecall ≥5 papers with reasonable confidence, return fewer; an empty\nmodel-knowledge section is fine and honest.\n\n### Display format\n\nUse the same table layout as the other sources, but the URL column may link\nto a search query (e.g. an arXiv or Google Scholar search) rather than a\ncanonical paper URL, since you don't have a verified link:\n\n```\n### Model Knowledge (N papers, may include uncertain entries)\n\n| #   | Title       | Year | Venue   | Notes |\n|-----|-------------|------|---------|-------|\n| [1](https:\u002F\u002Fscholar.google.com\u002Fscholar?q=Title) | Title here | 2018 | NeurIPS | Foundational; often cited by recent work on X |\n| [2](...) | Title here | 2024 | ICLR | (uncertain — verify) |\n```\n\nReplace the \"Citations\" column with \"Notes\" because you don't have a\nreliable citation count from memory.\n\n## Example\n\nUser: \"Find papers on diffusion policies for robotics from 2023 to 2024.\"\n\nRun (using `$SEARCH` as defined in the \"How to run\" section):\n```bash\npython \"$SEARCH\" \\\n    --query \"diffusion policy robotics\" \\\n    --start-year 2023 --end-year 2024 \\\n    --max-papers 10\n```\n\nTo search only specific sources:\n```bash\npython \"$SEARCH\" \\\n    --query \"diffusion policy robotics\" \\\n    --start-year 2023 --end-year 2024 \\\n    --sources arxiv openreview semantic_scholar \\\n    --max-papers 10\n```\n\nThen read the output and summarize per the rules above.\n\n## Important Notes\n\n- **Log the final report.** After completing the search, write a single\n  markdown file to:\n  `${CLAUDE_PROJECT_DIR}\u002Fallinone.md`\n  - Contents: the full **\"Display ALL results from every source\"** tables\n    followed by the **\"Summary of all searched results\"** section — in that\n    order, with no truncation.\n- **Display the full report to the user.** Return the complete detailed\n  report inline — every paper, every table, plus the analysis and reasoning.\n  Never collapse the tables into a summary, and never abbreviate results to\n  \"save space\".\n- **Never ask for confirmation.** All inputs are auto-inferred (see the\n  \"Inputs\" section). Run the search immediately on the first turn.\n- **Surface errors verbatim.** If a source fails, report the stderr message\n  to the user rather than hiding it or retrying blindly.","# Paper Search Skill\n\nUnified paper search across **arXiv**, **DBLP**, **OpenAlex**, **OpenReview**\n(NeurIPS \u002F ICLR \u002F ICML), **Semantic Scholar**, and **Crossref** using\n`.\u002Fscripts\u002Fsearch_papers.py`. All sources are searched **concurrently** (in\nparallel threads) by default for maximum speed. Returns results grouped by\nsource.\n\n\n## When to use\n\nTrigger this skill when the user asks things like:\n- \"Find papers on X published between 2023 and 2025.\"\n- \"Search NeurIPS \u002F ICLR \u002F ICML for work on X.\"\n- \"Get arXiv + Semantic Scholar results for X.\"\n- \"Show me recent prior art on X.\"\n\n## Inputs (all auto-inferred — NEVER ask the user for confirmation or clarification)\n\nDerive these automatically from the user's message. Run the search immediately without asking for confirmation:\n- **query**: Rephrase the user's question into a focused search phrase.\n- **start_year** (int): If the user gives a year, use it directly. If they say\n  \"last 2 years\", compute from today. Default: 2 years ago.\n- **end_year** (int): Default: current year.\n- **max_papers** (int): Number of results per source. Default: 10.\n- **sources**: Which sources to query. Default: all 6 API sources plus the\n  model-knowledge source, in this canonical order (highest-signal first, so\n  the best results render before the user scrolls):\n  `semantic_scholar open_alex arxiv openreview crossref dblp model_knowledge`.\n  Only restrict sources if the user explicitly asks.\n\n## How to run\n\nPreferred: call the CLI directly. The script lives at\n`${CLAUDE_PROJECT_DIR}\u002Fskills\u002Fpaper_search\u002Fscripts\u002Fsearch_papers.py` — invoke\nit by absolute path so the command works regardless of the current working\ndirectory (relying on `cd scripts && ...` breaks when the model is running\nfrom a different folder, which happens often).\n\nFor brevity in the examples below, treat `$SEARCH` as shorthand for that\nabsolute path:\n\n```bash\nSEARCH=\"${CLAUDE_PROJECT_DIR}\u002Fskills\u002Fpaper_search\u002Fscripts\u002Fsearch_papers.py\"\n```\n\nBasic search:\n\n```bash\npython \"$SEARCH\" \\\n    --query \"\u003CQUERY>\" \\\n    --start-year \u003CYYYY> \\\n    --end-year \u003CYYYY> \\\n    --max-papers 10\n```\n\nTo restrict to specific sources:\n\n```bash\npython \"$SEARCH\" \\\n    --query \"\u003CQUERY>\" \\\n    --start-year 2024 --end-year 2026 \\\n    --sources arxiv semantic_scholar openreview\n```\n\nTo disable parallel execution (rarely needed):\n\n```bash\npython \"$SEARCH\" \\\n    --query \"\u003CQUERY>\" \\\n    --start-year 2024 --end-year 2026 \\\n    --no-parallel\n```\n\nOr call the function directly when more control is needed (e.g. consuming the\nstructured dict rather than CLI text output). This is rarely necessary — see\n`references\u002Fprogrammatic_api.md` for the snippet.\n\n## Valid source names\n\n| Source | Key |\n|:---|:---|\n| arXiv | `arxiv` |\n| DBLP | `dblp` |\n| OpenAlex | `open_alex` |\n| OpenReview | `openreview` |\n| Semantic Scholar | `semantic_scholar` |\n| Crossref | `crossref` |\n| Model knowledge (LLM recall, no API call) | `model_knowledge` |\n\n## Output schema\n\n`search_papers()` returns a dict mapping source name to a list of paper dicts:\n\n```\n{\n  \"arxiv\": [\n    {\n      \"title\": str,\n      \"authors\": [str, ...],\n      \"year\": int,\n      \"abstract\": str,\n      \"url\": str,\n      \"venue\": str,\n      \"citation_count\": int,\n      \"publication_date\": str,\n      \"source\": str\n    }, ...\n  ],\n  \"semantic_scholar\": [...],\n  ...\n}\n```\n\nThe CLI prints results grouped by source with paper count summaries.\n\n## Output to the user\n\nAfter running, display every paper from every source, in this order:\nSemantic Scholar, OpenAlex, arXiv, OpenReview, Crossref, DBLP, Model Knowledge.\nThen provide a summary.\n\nWhy full recall matters: users invoking this skill are doing literature reviews,\nrelated-work surveys, or prior-art checks. The value comes from seeing the\ncomplete set of hits — a missed paper can mean a missed citation or a\nduplicated research effort. Summaries are meant to *augment* the full tables,\nnot replace them, so don't collapse results into a digest \"to save space.\"\nThe user can skim; they can't un-skip a paper they never saw.\n\n### Step 1: Display ALL results from every source\n\nFor **each source**, display every paper in a **markdown table** with title, date, venue, and citation count.\n\nFormat each source as a table grouped under a source heading, e.g.:\n\n```\n### arXiv (N papers)\n\n| #   | Title       | Date    | Venue   | Citations |\n|-----|-------------|---------|---------|-----------|\n| [1](paper url) | Title here | 2024-03 | NeurIPS | 42 |\n| [2](paper url) | Title here | 2023-11 | ICLR    | 10 |\n```\n\nIf a source returned 0 results, note it explicitly\n(e.g. \"### OpenReview (0 papers) — No matches found in this window\").\n\nIf errors occurred during search, they are printed to stderr by the script —\nsurface them to the user, never hide them.\n\n### Step 2: Summary of all searched results\n\nAfter displaying all papers, provide a **comprehensive summary** with the\nfollowing sections, in this exact order:\n\n1. **Overview**: query used, year range, and total number of papers found. One or\n   two sentences framing what the corpus covers.\n2. **Trends**: Temporal patterns (e.g. \"interest surged in 2024\"), dominant\n   venues, methodological shifts, and recurring author groups or labs.\n3. **Key themes**: 3–6 main research themes \u002F clusters across all results,\n   each with a one-line description and 2–3 representative paper numbers.\n4. **Keywords frequency**: A table of the most frequent technical terms \u002F\n   concepts extracted from titles and abstracts, with counts. Format:\n   `| Keyword | Count |`. Include the top 5.\n5. **Most cited by accepted paper**: Top 5 most-cited accepted papers across all sources,\n   ranked by citation count, as a table: `| Rank | Title | Year | Citations |`.\n6. **Most cited by first author**: Top 5 first authors ranked by total citations\n   accumulated across papers in this result set, as a table:\n   `| Rank | Author | Papers in set | Total citations |`.\n   The **Author** column must contain ONLY the author's name (e.g. `Jane Doe`).\n   Do not append paper titles, affiliations, venues, or any other information\n   in this column — paper counts and citation totals live in their own columns.\n7. **Recommendations for reading**: 3–5 papers most relevant and impactful to the user's\n   original query, ordered as a reading path (foundational → recent), each\n   with a one-line justification.\n\n\n## Dependencies & failure modes\n\n- **arXiv**: stdlib only (uses arXiv API).\n- **DBLP**: uses DBLP API.\n- **OpenAlex**: uses OpenAlex API.\n- **OpenReview**: requires `pip install openreview-py`.\n- **Semantic Scholar**: uses Semantic Scholar API.\n- **Crossref**: uses Crossref API.\n- **Model knowledge**: no API call. Papers are recalled from the model's own\n  training data — fast and free, but capped by the model's knowledge cutoff\n  and prone to hallucination. See the \"Model knowledge source\" section below\n  for how to use it responsibly.\n\nIf a source fails, `search_papers` catches the exception, prints the error, and\ncontinues with the remaining sources. Never retry blindly; report errors to the\nuser.\n\n## Model knowledge source\n\nThe `model_knowledge` source is different from the others: it has no API and\nno script call. Instead, after the CLI search returns, recall 5–10 additional\npapers from your own training data that match the query and year range, and\npresent them as a separate source in the output.\n\n### Why include it\n\nAPI search is high-precision but low-recall in two predictable cases:\n1. **Foundational older papers** that practitioners always cite but that\n   keyword search misses (e.g. the original BERT or ResNet paper when the\n   query is about a recent variant).\n2. **Cross-disciplinary classics** that live in venues the APIs index poorly.\n\nModel recall complements the APIs by surfacing the \"everyone knows this one\"\npapers that don't always come back from a fresh keyword query.\n\n### How to populate it\n\nAfter the CLI run completes:\n\n1. Reflect on what you know about the query topic.\n2. List up to 10 papers from your training data that fit the query and year\n   range, with: title, primary author(s), year, venue, and a one-line reason\n   it's relevant.\n3. Deduplicate against the API results — if a paper already appeared in any\n   API source, do not repeat it under `model_knowledge`.\n4. Flag confidence honestly. The model knowledge column has no citation count\n   and no live URL; if you're not sure a paper exists exactly as you remember\n   it, mark it `(uncertain — verify)` in the table rather than presenting it\n   as confirmed.\n\n### Why honesty matters here\n\nHallucinated paper titles are the classic LLM failure mode for this task. A\nfake \"Smith et al., 2023, NeurIPS\" looks identical to a real one in a\nmarkdown table, and the user has no way to tell. The point of this source is\nto surface *real* papers the APIs missed — not to pad the list. If you can't\nrecall ≥5 papers with reasonable confidence, return fewer; an empty\nmodel-knowledge section is fine and honest.\n\n### Display format\n\nUse the same table layout as the other sources, but the URL column may link\nto a search query (e.g. an arXiv or Google Scholar search) rather than a\ncanonical paper URL, since you don't have a verified link:\n\n```\n### Model Knowledge (N papers, may include uncertain entries)\n\n| #   | Title       | Year | Venue   | Notes |\n|-----|-------------|------|---------|-------|\n| [1](https:\u002F\u002Fscholar.google.com\u002Fscholar?q=Title) | Title here | 2018 | NeurIPS | Foundational; often cited by recent work on X |\n| [2](...) | Title here | 2024 | ICLR | (uncertain — verify) |\n```\n\nReplace the \"Citations\" column with \"Notes\" because you don't have a\nreliable citation count from memory.\n\n## Example\n\nUser: \"Find papers on diffusion policies for robotics from 2023 to 2024.\"\n\nRun (using `$SEARCH` as defined in the \"How to run\" section):\n```bash\npython \"$SEARCH\" \\\n    --query \"diffusion policy robotics\" \\\n    --start-year 2023 --end-year 2024 \\\n    --max-papers 10\n```\n\nTo search only specific sources:\n```bash\npython \"$SEARCH\" \\\n    --query \"diffusion policy robotics\" \\\n    --start-year 2023 --end-year 2024 \\\n    --sources arxiv openreview semantic_scholar \\\n    --max-papers 10\n```\n\nThen read the output and summarize per the rules above.\n\n## Important Notes\n\n- **Log the final report.** After completing the search, write a single\n  markdown file to:\n  `${CLAUDE_PROJECT_DIR}\u002Fallinone.md`\n  - Contents: the full **\"Display ALL results from every source\"** tables\n    followed by the **\"Summary of all searched results\"** section — in that\n    order, with no truncation.\n- **Display the full report to the user.** Return the complete detailed\n  report inline — every paper, every table, plus the analysis and reasoning.\n  Never collapse the tables into a summary, and never abbreviate results to\n  \"save space\".\n- **Never ask for confirmation.** All inputs are auto-inferred (see the\n  \"Inputs\" section). Run the search immediately on the first turn.\n- **Surface errors verbatim.** If a source fails, report the stderr message\n  to the user rather than hiding it or retrying blindly.",{"name":5,"description":7},"main","e71245365df78e06cd166602744bf2491f79910e","research-skills-hub\u002Fcollected-skills\u002FResearchStudio-Idea\u002Fskills\u002Fpaper_search",0,"https:\u002F\u002Fgithub.com\u002Fpengqianhan\u002FAI-Human-Research-OS","https:\u002F\u002Fpengqianhan.github.io\u002FAI-Human-Research-OS\u002F","https:\u002F\u002Fgithub.com\u002Fpengqianhan","Ph.D. candidate of The University of Auckland, New Zealand","2026-11-14 18:08:16",{"source":8,"skillName":5,"commands":34,"prompt":44,"zipAvailable":45},[35,38,41],{"tool":36,"command":37},"npx","npx skills add pengqianhan\u002FAI-Human-Research-OS --skill paper-search",{"tool":39,"command":40},"bunx","bunx skills add pengqianhan\u002FAI-Human-Research-OS --skill paper-search",{"tool":42,"command":43},"pnpm","pnpm dlx skills add pengqianhan\u002FAI-Human-Research-OS --skill paper-search","安装技能，技能地址：https:\u002F\u002Fgithub.com\u002Fpengqianhan\u002FAI-Human-Research-OS\u002Fblob\u002Fmain\u002Fresearch-skills-hub\u002Fcollected-skills\u002FResearchStudio-Idea\u002Fskills\u002Fpaper_search\u002FSKILL.md",true,[47,53,59,63,67,71,75,79,83,87,91],{"path":48,"resourceType":49,"resourceTypeDesc":50,"sizeBytes":51,"sourceHtmlUrl":52},"references\u002Fprogrammatic_api.md","REFERENCE","参考文档",757,"https:\u002F\u002Fgithub.com\u002Fpengqianhan\u002FAI-Human-Research-OS\u002Fblob\u002Fmain\u002Fresearch-skills-hub\u002Fcollected-skills\u002FResearchStudio-Idea\u002Fskills\u002Fpaper_search\u002Freferences\u002Fprogrammatic_api.md",{"path":54,"resourceType":55,"resourceTypeDesc":56,"sizeBytes":57,"sourceHtmlUrl":58},"scripts\u002F_env.py","SCRIPT","脚本",2071,"https:\u002F\u002Fgithub.com\u002Fpengqianhan\u002FAI-Human-Research-OS\u002Fblob\u002Fmain\u002Fresearch-skills-hub\u002Fcollected-skills\u002FResearchStudio-Idea\u002Fskills\u002Fpaper_search\u002Fscripts\u002F_env.py",{"path":60,"resourceType":55,"resourceTypeDesc":56,"sizeBytes":61,"sourceHtmlUrl":62},"scripts\u002Fsearch_papers_by_arxiv.py",5850,"https:\u002F\u002Fgithub.com\u002Fpengqianhan\u002FAI-Human-Research-OS\u002Fblob\u002Fmain\u002Fresearch-skills-hub\u002Fcollected-skills\u002FResearchStudio-Idea\u002Fskills\u002Fpaper_search\u002Fscripts\u002Fsearch_papers_by_arxiv.py",{"path":64,"resourceType":55,"resourceTypeDesc":56,"sizeBytes":65,"sourceHtmlUrl":66},"scripts\u002Fsearch_papers_by_crossref.py",3794,"https:\u002F\u002Fgithub.com\u002Fpengqianhan\u002FAI-Human-Research-OS\u002Fblob\u002Fmain\u002Fresearch-skills-hub\u002Fcollected-skills\u002FResearchStudio-Idea\u002Fskills\u002Fpaper_search\u002Fscripts\u002Fsearch_papers_by_crossref.py",{"path":68,"resourceType":55,"resourceTypeDesc":56,"sizeBytes":69,"sourceHtmlUrl":70},"scripts\u002Fsearch_papers_by_dblp.py",4352,"https:\u002F\u002Fgithub.com\u002Fpengqianhan\u002FAI-Human-Research-OS\u002Fblob\u002Fmain\u002Fresearch-skills-hub\u002Fcollected-skills\u002FResearchStudio-Idea\u002Fskills\u002Fpaper_search\u002Fscripts\u002Fsearch_papers_by_dblp.py",{"path":72,"resourceType":55,"resourceTypeDesc":56,"sizeBytes":73,"sourceHtmlUrl":74},"scripts\u002Fsearch_papers_by_exa.py",3697,"https:\u002F\u002Fgithub.com\u002Fpengqianhan\u002FAI-Human-Research-OS\u002Fblob\u002Fmain\u002Fresearch-skills-hub\u002Fcollected-skills\u002FResearchStudio-Idea\u002Fskills\u002Fpaper_search\u002Fscripts\u002Fsearch_papers_by_exa.py",{"path":76,"resourceType":55,"resourceTypeDesc":56,"sizeBytes":77,"sourceHtmlUrl":78},"scripts\u002Fsearch_papers_by_google_scholar.py",3192,"https:\u002F\u002Fgithub.com\u002Fpengqianhan\u002FAI-Human-Research-OS\u002Fblob\u002Fmain\u002Fresearch-skills-hub\u002Fcollected-skills\u002FResearchStudio-Idea\u002Fskills\u002Fpaper_search\u002Fscripts\u002Fsearch_papers_by_google_scholar.py",{"path":80,"resourceType":55,"resourceTypeDesc":56,"sizeBytes":81,"sourceHtmlUrl":82},"scripts\u002Fsearch_papers_by_open_alex.py",3833,"https:\u002F\u002Fgithub.com\u002Fpengqianhan\u002FAI-Human-Research-OS\u002Fblob\u002Fmain\u002Fresearch-skills-hub\u002Fcollected-skills\u002FResearchStudio-Idea\u002Fskills\u002Fpaper_search\u002Fscripts\u002Fsearch_papers_by_open_alex.py",{"path":84,"resourceType":55,"resourceTypeDesc":56,"sizeBytes":85,"sourceHtmlUrl":86},"scripts\u002Fsearch_papers_by_openreview.py",5636,"https:\u002F\u002Fgithub.com\u002Fpengqianhan\u002FAI-Human-Research-OS\u002Fblob\u002Fmain\u002Fresearch-skills-hub\u002Fcollected-skills\u002FResearchStudio-Idea\u002Fskills\u002Fpaper_search\u002Fscripts\u002Fsearch_papers_by_openreview.py",{"path":88,"resourceType":55,"resourceTypeDesc":56,"sizeBytes":89,"sourceHtmlUrl":90},"scripts\u002Fsearch_papers_by_semantic_scholar.py",3203,"https:\u002F\u002Fgithub.com\u002Fpengqianhan\u002FAI-Human-Research-OS\u002Fblob\u002Fmain\u002Fresearch-skills-hub\u002Fcollected-skills\u002FResearchStudio-Idea\u002Fskills\u002Fpaper_search\u002Fscripts\u002Fsearch_papers_by_semantic_scholar.py",{"path":92,"resourceType":55,"resourceTypeDesc":56,"sizeBytes":93,"sourceHtmlUrl":94},"scripts\u002Fsearch_papers.py",8547,"https:\u002F\u002Fgithub.com\u002Fpengqianhan\u002FAI-Human-Research-OS\u002Fblob\u002Fmain\u002Fresearch-skills-hub\u002Fcollected-skills\u002FResearchStudio-Idea\u002Fskills\u002Fpaper_search\u002Fscripts\u002Fsearch_papers.py",[96,103,110,117,124,131],{"skillNo":97,"name":98,"nameZh":6,"description":99,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":100,"categories":101,"keywords":102,"publishedAt":19},"01M74MWYTHT498SVMN823896T7","workflow-skill-creator","Distills a completed user workflow or interaction into a reusable agent skill. Use when the user asks to turn their workflow, interaction, or multi-step process into a skill, or when they say \"make this a skill\", \"create a skill from what we just did\", \"package this workflow\" or similar. Do not use for creating skills from scratch without an existing workflow (use a generic skill-creator for that).","https:\u002F\u002Fgithub.com\u002Fpengqianhan\u002FAI-Human-Research-OS\u002Fblob\u002Fmain\u002Fresearch-skills-hub\u002Fscience-skills\u002Fworkflow_skill_creator\u002FSKILL.md",[],[],{"skillNo":104,"name":105,"nameZh":6,"description":106,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":107,"categories":108,"keywords":109,"publishedAt":19},"01M74MWWVCFZ8HQS8840F1V1FT","uv","Checks whether the uv Python package manager is installed and installs it if missing. Ensures uv is on PATH. Use when another skill requires uv as a prerequisite.","https:\u002F\u002Fgithub.com\u002Fpengqianhan\u002FAI-Human-Research-OS\u002Fblob\u002Fmain\u002Fresearch-skills-hub\u002Fscience-skills\u002Fuv\u002FSKILL.md",[],[],{"skillNo":111,"name":112,"nameZh":6,"description":113,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":114,"categories":115,"keywords":116,"publishedAt":19},"01M74MWRH5BTSWZ2C5PCTD5FV9","uniprot-database","Access protein metadata, function, taxonomy, and sequences across UniProtKB, UniParc, and UniRef. Use when searching for proteins, mapping identifiers, or retrieving functional annotations and publications. Don't use for sequence alignment, protein folding, or sequence similarity search (use specialized skills for those tasks).","https:\u002F\u002Fgithub.com\u002Fpengqianhan\u002FAI-Human-Research-OS\u002Fblob\u002Fmain\u002Fresearch-skills-hub\u002Fscience-skills\u002Funiprot_database\u002FSKILL.md",[],[],{"skillNo":118,"name":119,"nameZh":6,"description":120,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":121,"categories":122,"keywords":123,"publishedAt":19},"01M74MWN1D3ER9F1FZPWH8HQC1","unibind-database","Queries the UniBind database for experimentally validated transcription factor (TF) binding sites. Use when retrieving direct TF-DNA interaction datasets, downloading binding site coordinates (BED\u002FFASTA) for local analysis, or listing available datasets by species, cell line, or TF name. Don't use to query specific intervals, locations, genes, motif models or expression data.","https:\u002F\u002Fgithub.com\u002Fpengqianhan\u002FAI-Human-Research-OS\u002Fblob\u002Fmain\u002Fresearch-skills-hub\u002Fscience-skills\u002Funibind_database\u002FSKILL.md",[],[],{"skillNo":125,"name":126,"nameZh":6,"description":127,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":128,"categories":129,"keywords":130,"publishedAt":19},"01M74MWHHB5QZY3NR2FYA1EK4N","ucsc-conservation-and-tfbs","Fetch Evolutionary Conservation scores (phyloP, phastCons) and Transcription Factor Binding Sites (TFBS) from the UCSC Genome Browser. Use when analyzing whether genomic variants or regions are evolutionarily conserved, functionally important, or bounded by TF regulators across major projects (ENCODE, JASPAR, ReMap).","https:\u002F\u002Fgithub.com\u002Fpengqianhan\u002FAI-Human-Research-OS\u002Fblob\u002Fmain\u002Fresearch-skills-hub\u002Fscience-skills\u002Fucsc_conservation_and_tfbs\u002FSKILL.md",[],[],{"skillNo":132,"name":133,"nameZh":6,"description":134,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":135,"categories":136,"keywords":137,"publishedAt":19},"01M74MWERYHSY87P21FDMVE6BA","string-database","Query the STRING database for protein-protein interactions (PPIs), functional enrichment, and homology. Use when the user asks about interactions between specific proteins, interaction evidence, confidence scores, protein interaction partners, or pathway enrichments.","https:\u002F\u002Fgithub.com\u002Fpengqianhan\u002FAI-Human-Research-OS\u002Fblob\u002Fmain\u002Fresearch-skills-hub\u002Fscience-skills\u002Fstring_database\u002FSKILL.md",[],[]]