180 個技能
Autonomous Goal-directed Iteration. Apply Karpathy's autoresearch principles to ANY task. Loops autonomously — modify, verify, keep/discard, repeat. 9 subcommands: plan, debug, fix, security, ship, scenario, predict, learn.
CHARLS (China Health and Retirement Longitudinal Study) database-specific knowledge for reproducing published papers. Use when reproducing or analyzing papers that use CHARLS data, including variable mapping from harmonized to raw questionnaire items, cognitive function scoring (episodic memory, mental status, TICS), CESD-10 depression screening, social isolation index construction, and chronic disease coding. Also use for any CHARLS data cleaning, variable construction, or cohort selection task.
Standard operating procedure for research survey and project execution. Covers deep reading methodology, information synthesis, paper evaluation, citation integrity, cross-module coordination (Library↔Tasks↔Workspace↔Monitor), and research project lifecycle (Phase 2-4).
Semantic search across UK Biobank's 12,000+ data fields and publications — find the right variables for your research question.
Research methodology guide. Referenced by the research-designer and statistical-analyst agents when selecting research designs and analysis methods. Use for 'research methodology', 'experiment design', or 'statistical analysis' requests. Actual data collection and IRB review processing are out of scope.
Develop survey instruments, implement sampling strategies, optimize response rates, and manage multi-mode data collection
Use when `evidence-review` needs an operational protocol before screening and extraction. **Trigger**: protocol, PRISMA, systematic review, inclusion/exclusion, 检索式, 纳入排除. **Use when**: `evidence-review` pipeline 的起点（C1），需要先锁定 protocol 再开始 screening/extraction。 **Skip if**: 不是做 evidence/systematic review（或 protocol 已经锁定且不允许修改）。 **Network**: none. **Guardrail**: protocol 必须包含可执行的检索与筛选规则；需要 HUMAN 签字后才能进入 screening。
Research output templates, distillation workflow, and quality standards for evidence-driven research
Source reputation tiers, cross-referencing methodology, bias detection, and citation format requirements
Generates high-reliability research questionnaires using the 5-point Likert scale based on objectives, and transforms binary questions into agreement scales.
Analyze a user's stated beliefs regarding a specific statement to determine their stance and recommend the best matching option from a standard 5-point Likert scale.
Conduct comprehensive deep research on any topic using Dify-powered workflow - searches documentation, academic papers, tutorials, APIs, best practices, and returns structured analysis with insights.
Research analyst in a multi-bot group chat — synthesizes background information, fact-checks claims, identifies knowledge gaps, and suggests next steps. Use when the conversation needs context, a factual claim needs verification, someone asks for background research, or a discussion lacks supporting evidence.
执行完整的 7 阶段深度研究流程。接收结构化研究任务，自动部署多个并行研究智能体，生成带完整引用的综合研究报告。当用户有结构化的研究提示词时使用此技能。
将原始研究问题细化为结构化的深度研究任务。通过提问澄清需求，生成符合 OpenAI/Google Deep Research 标准的结构化提示词。当用户提出研究问题、需要帮助定义研究范围、或想要生成结构化研究提示词时使用此技能。
Sibyl 怀疑论者 agent - 以最大怀疑态度审视实验结果
Sibyl 反思 agent - 分析迭代产出，分类问题，生成改进计划
Sibyl 乐观分析者 agent - 从积极角度分析实验结果
Sibyl pilot 验证决策 agent - 根据小实验结果决定 ADVANCE / REFINE / PIVOT
Sibyl 实验主义者 agent - 关注可复现性、数据质量和实验设计严谨性
Sibyl 监督决策 agent - 分析实验结果决定 PIVOT 还是 PROCEED
Sibyl 结果辩论综合者 agent - 综合6方结果分析形成统一判断和行动计划
Best practices for designing reproducible experiments
6 social science skills. Trigger: survey research, social networks, psychology, behavioral studies. Design: quantitative and qualitative methods for social science research.
Conduct qualitative meta-synthesis and evidence synthesis methods
Systematic review methodology with PRISMA and evidence synthesis
Search papers and analyze citation graphs via OpenAlex and CrossRef APIs
Bulk download PMC Open Access articles via FTP for large-scale mining
Deposit and discover research datasets via Harvard Dataverse API
Questionnaire and survey design with Likert scales and coding