TerryFYL

研究方向发现引擎。在成熟领域的交界处发现没人建的桥。 触发条件: "找方向"/"研究方向"/"选题"/"下一轮做什么"/"research frontier"。 核心方法: 资产盘点 → 结构选型 → 交叉缝隙扫描(deep research驱动) → 证据验证 → 可行性确认。

AI 科研军团统一入口。用户投入任何科研任务→Wei 读取任务→查阅 Skill 注册表→动态调度各角色。 触发条件: 用户说"/start-army"、"启动军团"、"接单"、"开始项目"，或提供数据文件+研究需求。

原始数据真伪与异常模式扫描。既支持 /data-forensics，也支持自然语言触发，如‘这份数据可信吗’。

数据探查与画像。30分钟生成完整数据画像，包括变量字典、分布、缺失率。既支持 /data-profiler，也支持自然语言触发，如‘先帮我看看这份数据长什么样’。

最终交付收口。把投稿包整理成可交付目录，补齐说明文件、清单和接收指引。既支持 /delivery，也支持自然语言触发，如‘把这些整理成最终交付目录’。

学术论文撰写。基于统计结果和文献生成IMRAD格式论文。既支持 /manuscript-draft，也支持自然语言触发，如‘帮我写成论文初稿’。

质量审查与自动迭代。多轮审查直到通过或达到上限。既支持 /quality-review，也支持自然语言触发，如‘帮我审这篇稿子’。

研究设计与方案规划。从数据字典出发，设计研究问题、分析策略和叙事线。既支持 /research-design，也支持自然语言触发，如‘帮我设计研究方案’。

Audit supplied research data for identity, provenance, schema, missingness, duplicates, impossible values, temporal and unit consistency, leakage, and transformation traceability. Use before analysis to determine data readiness without claiming to detect fabrication.

Build a traceable evidence foundation for research claims from user-provided sources and retrieval tools that are actually available. Use for literature grounding, source identity checks, claim-evidence mapping, evidence gaps, and creation of the canonical evidence-ledger.json sidecar.

Stress-test a manuscript by mapping supplied evidence to scientific concerns, consequences, and revision locations. Use for structured scientific critique, aspect-coverage audits, unresolved-concern tracking, and fatal-major-minor prioritization without impersonating an editor or reproducing peer-review corpora.

Create an editable evidence-bounded manuscript draft from supplied research material and the journal pack selected by the Pi extension. Use for journal-shaped section planning and drafting while preserving missing evidence as explicit placeholders and keeping current author-instruction checks unresolved.

Triage a manuscript and supplied submission files for package completeness, declarations, claim-to-evidence traceability, and prioritized blockers. Use before pre-review or packaging when a manuscript, excerpt, file inventory, evidence ledger, or journal requirements are available.

Run an evidence-first pre-submission review across design, data, methods, statistics, results, claims, reproducibility, ethics, and journal-pack rules. Use after manuscript preflight to create a prioritized revision board without predicting editorial acceptance.

Convert an editor decision and reviewer comments into a complete traceable revision ledger, evidence-bounded response draft, and verified manuscript change log. Use for major or minor revision, point-by-point responses, resubmission planning, and revision coverage audits.

Convert a qualified research brief and evidence ledger into a testable study-design contract with an explicit estimand, measurement plan, bias analysis, analysis plan, falsifiers, and governance gates. Use after topic qualification and before data analysis or manuscript drafting.

Statistical analysis service. Trigger conditions: (1) Upload a data file, (2) Say "statistical analysis" / "analyze my data" / "run the data", (3) Mention a specific statistical method (t-test / regression / SEM, etc.). **Core principle**: Proactively diagnose like a top-tier statistical consultant — don't just execute the method the user requests, ensure the method selection is correct.

幻灯片/PPT生成器。支持从学术PDF文献自动提取图表并生成汇报PPT。前置一次性素材收集，之后全程自主执行，不打断用户。学术文献模式使用故事线驱动的python-pptx原生渲染。v4.1新增图表矢量重绘：从栅格图中用AI提取数据点，matplotlib重绘为真矢量SVG。