Research Claude Skills
2 269 research skills from 396 authors. They find sources and get you up to speed on unfamiliar ground. Half of them fit into 2 278 tokens or less — that is what one costs your context window when the agent loads it. 663 ship runnable scripts rather than instructions alone. 4 of them cannot work without an MCP server, most often rube. We also found 264 copies of these same skills sitting in other people's repositories — counted once here, not 264 times.
2 269 unique 396 authors 1 169 updated this month 93 from vendors
Use whenever the user wants to obtain, download, or fetch a paper's PDF — given a DOI, an arXiv id, a paper title, a citation, or a list of DOIs. Trigger on phrases like "download this paper", "find the PDF for [DOI]", "grab me the [Nature/bioRxiv/arXiv] paper on X", "get the open-access version", "I need this article", or any bulk/batch paper download request, even when the user doesn't explicitly say "PDF" or "DOI". Resolves via Unpaywall → Semantic Scholar → arXiv → PubMed Central → bioRxiv/medRxiv → publisher direct (institutional opt-in) → Sci-Hub mirrors as last-resort fallback.
Suede-owned customer-research discipline for interview design, transcript and ticket synthesis, review and forum mining, quote banks, jobs, and evidence-backed personas. Use when discovering or synthesizing what a defined customer segment actually says, does, needs, and resists. NOT FOR: competitor-only profiling (use suede-competitor-profiling), writing final marketing copy (use suede-copy), or deciding product priorities without product evidence (use suede-product-marketing).
Suede-affiliated directory distribution strategy for selecting listings, sequencing submissions, tailoring positioning, and verifying backlinks. Use when a product needs startup, SaaS, AI, MCP, marketplace, or review-directory submissions and a measurable tracker. NOT FOR: broader launch orchestration (use suede-launch-packaging), scalable destination-page production (use suede-programmatic-seo), or citation and search auditing (use suede-seo-audit).
Grade a public page for launch appeal: findability, first-screen clarity, CTA pull, proof quality, and AI citation readiness.
Parallel persona planning for new projects. Research agent runs first to build domain context, then Architect, PM, and Security agents run in parallel. Synthesis agent combines all perspectives into a detailed GSD-style PLAN.md with Tensions section.
USE FOR web search, research, RAG, grounding, browse, find, lookups, fact-checking, documentation, agentic AI. All-in-one, optimized for AI agents. Pre-extracted, token-budgeted web content, deep research, news, images, videos, places, custom ranking
USE FOR AI-grounded answers via OpenAI-compatible /chat/completions. Two modes: single-search (fast) or deep research (enable_research=true, thorough multi-search). Streaming/blocking. Citations.
Use when the user wants a complete, publication-grade research paper on a specific topic — produces 200+ real citations, 4–8 publication-grade figures, and 7 sections of substantive prose compiled to PDF in one pass. No skeleton stage.
Use when the user wants an end-to-end AI4S research pipeline — broad direction or specific topic in, full research package out (exploration + literature survey + experiment + paper). Meta-skill that chains the four downstream skills in order. Pure markdown, no Python runtime.
Use when the user has a vague research direction and wants to explore feasible specific topics. Outputs a structured analysis with candidate topics, innovation/feasibility scoring, and a pre-survey of 20–30 representative works. Single-stage, no Python runtime.
Use when the user wants a comprehensive literature survey on a specific research topic. Outputs a complete PDF survey (6–20 pages, 60+ real citations, 100+ recommended) with LaTeX source, topic-specific publication figures, and a classified literature table. Single-stage, no Python runtime.
Use when the user wants a paper audited for integrity issues — image misuse, numerical anomalies, logical gaps — and needs a reviewable evidence report. Works on external papers (PDF / DOI / arXiv) and on outputs from a local paper-writer run. Single-stage skill.
在 Codex App 中为复杂、可并行的知识工作或编程任务自动创建多个可指定模型与推理强度的后台任务，由主 Agent 负责规划、分工、集成和验收。用于多来源调研、多章节内容、复杂 Skill/PPT、跨模块开发、独立验证或 2 个以上互不依赖工作流；也用于用户明确要求模型路由、后台 Worker、Agents Team 或持久项目协作。简单问答、状态查询、单文件小改、强顺序任务和发布/付款/删除/账户操作不得自动触发。
> Evidence-based investigation of issues, bugs, API calls, integrations, and other aspects of software development that need a deep dive to find the root cause and solutions. Use when you need to debug, troubleshoot, diagnose, or figure out why something is broken. Does not review code for quality or style — use code-review for auditing changes or post-code-review-to-pr for posting review feedback to GitHub. Does not assess architectural health or structural risk — use architectural-analysis for architectural concerns. Does not research open-ended options, prior art, or how something works when nothing is broken — use research for that. Does not capture feedback on Han's own skills — use han-feedback for that.
> Performs a gap analysis between two artifacts (a current state and a desired state) and produces a plain-language, stakeholder-readable report indexed by stable gap IDs. Use when the user wants to compare, evaluate, audit, or reconcile one artifact against another. Does not investigate runtime bugs — use investigate. Does not assess module-level architecture — use architectural-analysis. Does not research open-ended options with no second artifact to compare against — use research.
Researches an open-ended question — options, possible solutions, prior art, trade-offs, or how something works — and produces a durable, evidence-backed, adversarially-validated report that recommends an option without committing the team to any artifact. Use when you want to research approaches, weigh options, survey prior art or the state of the art, or understand how something works before committing to a direction. Does not diagnose a bug, failure, or root cause — use investigate. Does not specify a feature — use plan-a-feature. Does not create or update a coding standard — use coding-standard. Does not compare two concrete artifacts for gaps — use gap-analysis. Does not assess an existing module's architecture — use architectural-analysis. Does not capture feedback on Han's own skills — use han-feedback.
> Quality control metrics and filtering thresholds for protein design. (2) Setting filtering thresholds for pLDDT, ipTM, PAE, (3) Checking sequence liabilities (cysteines, deamidation, polybasic clusters), (4) Creating multi-stage filtering pipelines, (5) Computing PyRosetta interface metrics (dG, SC, dSASA), (6) Checking biophysical properties (instability, GRAVY, pI), (7) Ranking designs with composite scoring. This skill provides research-backed thresholds from binder design competitions and published benchmarks.
Research external context for a dataset — domain background, history, related studies, and why this data matters. Outputs detective.json (structured findings with det_xx IDs) before any analysis begins.
Research external context for a dataset — domain background, history, related studies, and why this data matters. Outputs detective.json (structured findings with det_xx IDs) before any analysis begins.
专用于小说写作的结构化深度研究 Skill。支持对题材趋势、平台生态、专业知识、场景环境、人物原型、写作技法、读者市场等研究域，执行大纲生成→并行深搜→报告输出的全流程深度调研。关键词：深度研究、写作研究、题材调研。
向现有小说写作调研outline补充字段定义。优先补充对创作者有直接价值的"创作可迁移"维度字段。
读取调研outline，为每个item启动独立agent，从小说创作者视角进行深度调研。禁用task output。
Add field definitions to existing research outline.
Add items (research objects) to existing research outline.
向现有调研outline补充items（调研对象）。
Read research outline, launch independent agent for each item for deep research. Disable task output.
Summarize deep research results into markdown report, cover all fields, skip uncertain values.
Conduct preliminary research on a topic and generate research outline. For academic research, benchmark research, technology selection, etc.
将 deep 调研结果汇总为小说创作者可直接使用的 Markdown 研究报告，覆盖所有字段，跳过不确定值。报告落盘到项目或题材的 `写作研究/` 目录。
对小说写作相关话题进行初步调研，生成调研outline。用于题材趋势、平台生态、专业知识、场景环境、人物原型、写作技法、读者市场等研究域的深度调研。
用于从写作研究报告（题材×平台级）中提取可执行写作约束，产出"写作研究模板"，作为下游写/审/改 Skill 的题材×平台默认约束基线。蒸馏源默认扫描小说项目根"写作研究/"目录 + "调研报告/"目录 + "CommonSkills/写作研究/"目录；支持指定目录、文件、URL 或 网络搜索内容。关键词：蒸馏写作研究、研究报告提取、题材平台约束、写作约束基线、规则提取。
设计、重写或补强故事设定全品类——包括世界观设定、技术设定、规则/机制设定、社会制度设定、能力/金手指设定、手法/程序链/证据载体设定。用于产出可叙事化、可证据化、边界与代价明确的作者侧设定裁判源，覆盖规则条款、误导与证伪、程序摩擦、镜头包、实际文件写回与跨阶段演进。支持联动竞对分析、网络调研、深度研究与结构化深入思考，确保设定既有真实感锚点又有竞争辨识度。适合所有品类故事设定的创建、重构、补强与规则边界修复。
创建异能志怪题材章节正文、作者有话说与章节后记。用于按分卷大纲、人物传记、设定与写作研究，直接新写、重写或扩写都市异术超能 / 志怪连载章节；作为题材包装层与路由层，负责保留异能志怪标准入口名，并要求生成的 `
| Apply Chain-of-Verification (CoVe) prompting to improve response accuracy through self-verification. Use when complex questions require fact-checking, technical accuracy, or multi-step reasoning.
| Apply Chain-of-Verification (CoVe) prompting to improve response accuracy through self-verification. Use when complex questions require fact-checking, technical accuracy, or multi-step reasoning.
> Use when the user wants an autonomous ML research loop that explores the space broadly rather than several wild, diverse swings (full rewrites, different architectures/regimes) early, then enters an adaptive phase that picks swing / merge / exploit per iteration — with a hard stagnation guard that bans further small-step exploits once they run too long, forcing a pivot back to a swing or merge. Tracks an approaches.md registry and a move_type per iteration; analyses every run before the next move. One change per iteration; loops forever until interrupted. Not for the standard analysis-first ml-autoresearch (which lets analysis alone choose each change), one-off training runs, or sweeps.
> Use when the user wants to generate and literature-vet a pool of novel, testable research hypotheses LiteratureScout grounds each in real retrieved literature (already known? closest prior work? what gap does it fill?), and a Judge scores them against a fixed rubric and keeps the strong, non-duplicate ones; rounds repeat — mutating toward the open gaps — until fresh rounds stop adding keepers. Not for sharpening or decomposing a research question (no grounding/scoring there), and not for grading an existing written proposal against the literature.
> Use when the user wants two approaches raced head-to-head on a single shared metric — e.g. a classical/algorithmic lane vs an ML/learned lane, or any two strategies for the same task. Each lane runs its own analysis-first research loop confined to its lane, the lanes share a scoreboard and may borrow ideas across the boundary without abandoning their identity, and a shared eval keeps the head-to-head honest; loops until interrupted, reporting the current leader. Not for improving a single approach in isolation (use a single-track research loop), and not for picking between two finished artifacts in one shot (that is a one-time comparison).
> Use when a loop needs scholarly literature — paper discovery, novelty checks, full-text snippet search, citation-graph traversal, single-paper reads, or experimental-result extraction. A shared, stdlib-only CLI (`tools/lit_search.py`) over Semantic Scholar + arXiv (plus optional OpenAlex, of them call it instead of vendoring their own copy. Degrades to the caller's WebSearch/WebFetch.
> Use when the user wants a structured, saturating literature survey on a question — not a one-shot summary, but an evidence/contradiction matrix (sources × claims) built by iterative search until coverage stops growing. Each round expands the search (new sub-topic queries plus citation-graph walks), admits new sources, extracts their claims with verbatim snippets, and records where every source stands on each claim (supports / contradicts / qualifies); the feedback signal is how many new matrix-changing sources a round adds, and it stops at saturation, patience, or budget. The output is the matrix plus a synthesis of consensus, disputes, and gaps, every cell backed by a real citation. Not for grading a written proposal against the literature (that is a proposal-evaluation task) and not for generating new hypotheses — this maps what the literature already says.
> Use when the user wants an autonomous ML research loop that does more than blindly try changes. After every training run the agent analyses what actually happened inside the model — gradients, activations, embeddings, errors, data — and grounds the next change in that evidence. A `<literature>` searches papers, grades the evidence, and implements only what prior work supports. One change per run; loops forever until interrupted. Not for one-off training runs or hyperparameter sweeps.
> Use when the user has a research proposal (problem + proposed methodology + planned experiments) and wants it iteratively strengthened until it clears a passing grade. ScholarEval grades the proposal against the literature (Soundness + Contribution), a Judge scores that feedback 0-100 on a fixed rubric, and a Reviser rewrites the proposal to fix the worst points without diluting the research question; loops until the grade passes or the budget is hit. Not for generating a proposal from scratch, and not for running a literature survey on its own.
> Use when the user has a scientific draft (with its dataset, figures, and optional analysis code) and wants it iteratively revised until it clears a quality bar. Five specialist judges (figures, scientific content, style, formatting, code) critique the draft; a fresh, independent peer_reviewer grades it on those same axes (1-5 each → a percentage) with honesty guardrails; a scientific_writer revises prose, figures, and code — regenerating figures by running the user's plot command — until the peer-review score clears the threshold or the budget is hit. Not for writing a paper from a blank page, and not for a standalone literature survey.
> Use when the user has a vague topic or area of interest and wants it sharpened into a few strong, novel, feasible research questions. Drafts candidate questions, scores each against a fixed rubric (Specific, Answerable, Novel, Feasible, Significant) with a light literature/web novelty check, and revises the weakest axis until enough questions clear the bar. Not for grading a full written proposal (use the research-proposal loop), and not for turning a question into testable predictions (use the hypothesis-generation loop).
> Use when the user has scientific data (or a prompt alluding to scientific data) and wants a publication-quality figure made from it. A generator drafts and renders a figure that lands a frozen communication goal; an adversarial critic critiques it hard and grades it 1-5 per axis against a fixed rubric (message, aesthetic, clarity, integrity, and a conditional domain-completeness axis), aggregates to 0-100, and decides pass; the generator revises against the critic's findings until the grade clears a threshold or the budget is hit. Both roles may consult the literature (Semantic Scholar + arXiv) to verify domain content (e.g. a pathway figure's gene set, or a benchmark's reported numbers), and conform to / grade against a named journal's figure spec fetched via web search. Not for writing a paper or analyzing a dataset, and not for editing an existing finished image — this renders a figure from data/brief and iterates on it.
| 当系统提示需要设计引用格式、信息溯源机制、来源标注系统时调用。适用于文档问答、搜索增强生成（RAG）、代码引用、浏览器辅助等需要让用户追溯信息来源的场景。不适用于纯创作类输出（如故事、诗歌），不适用于无需溯源的常识问答，也不适用于注入防御（虽然两者都涉及内容可信度）。
> End-of-milestone audit of a research project's `Code/` folder against the paper and outputs. Surfaces drift between scripts, YAML headers, the script registry, outputs, raw data, and the paper. Categorises findings as MUST-FIX or SUGGESTED and proposes fixes one at a time for the user to confirm. Use this skill when the user says "code sweep", "sweep the code", "run a code sweep", "audit my code folder", "check code freshness", "is everything up to date", "any stale outputs", "audit code consistency", or similar — typically before a paper submission, milestone, or share. Extends `script-registry`; builder rather than redefining either.
> Per-project glossary of key definitions, abbreviations, and command-phrases, stored in `GLOSSARY.md` at the project root. Use this skill when the user defines or asks about a project-specific term — variable names, dataset or database names, acronyms — or sets up a command-phrase (a phrase that maps to an action, e.g. "push" = commit and push the paper to GitHub). Triggers "what does X mean here", "what does X stand for", "from now on X means Y", "show the glossary", "what's in our glossary", and "remove X from glossary". passing, or when you hit an undefined abbreviation or variable name in their code or data. Loaded at session start by `/spin-up` so command-phrases stay active.