Menu
Skills in this repository
SkillsMP has collected 60 skills from MengkunLiang/DIG-ResearchOS. Open a skill to review its source and details.
MengkunLiang/DIG-ResearchOSShowing 20 of 60 collected skills.
Answer a bounded research question from exact sections of one uploaded paper PDF, preserving page and section anchors and evidence boundaries. Use when a researcher needs a precise method, result, limitation, dataset, or claim check without a full literature…
Turn one uploaded or explicitly identified research paper into a source-traceable standalone paper reading note with clear locations in the paper. DOI, arXiv ID, OpenAlex ID, or a direct URL can be resolved into the declared input PDF path before reading.
从 T5 reboost 或 legacy handoff 及其权威来源构建、预览或校验项目专属 external-executor Skill Suite。适用于 Suite 缺失、过期、恢复前核验或发布失败诊断；不会补造项目事实、执行实验或替换运行中的 Suite。
Build a source-grounded Related Work section package from a target paper position, literature artifacts, and verified bibliography. It produces a narrative section draft plus citation and evidence audits without inventing citations or positioning claims. Use…
Produce a factual research-landscape report from verified literature artifacts, including domain structure, method families, evidence coverage, tensions, resource availability, and evidence-bounded research opportunities. Use when a researcher needs a…
Register user-provided research PDFs, datasets, code, URLs, and context notes into a bounded, source-traceable ResearchOS material manifest. Use before literature work, experiment design, paper writing, or a standalone analysis when a researcher needs to…
Prepare a human-approved, evidence-audited survey package from a domain synthesis before any Survey prose is drafted. It checks corpus sufficiency, proposes taxonomy and storyline, identifies weak classes, and can request or perform scoped literature…
Review a LaTeX manuscript against a user-supplied conference or journal scope, format, and evaluation expectations, producing a source-aware venue-fit and revision-priority report without changing the manuscript. Use before choosing a venue, polishing a…
Audit whether manuscript citations and claim wording are traceable to available BibTeX and paper-note evidence. Use before paper polishing, revision, or submission when citation hallucination risk must be made explicit.
Review a proposed research experiment for identification, controls, measurements, baseline fairness, feasibility, and stopping criteria. Use before external execution; this Skill never runs code or changes an executor project.
Build a transparent, multilingual and reproducible literature query portfolio from a scoped research brief. Use before web search when the researcher needs explicit source coverage, language policy, inclusion logic, and targeted evidence gaps.
Identify concrete datasets, benchmarks, baselines, code artifacts, and reproducibility constraints from a research brief and existing literature artifacts. Use when a researcher needs a resource feasibility report without treating unverified web snippets as…
Audit manuscript numeric and strength-of-claim statements against a normalized ResearchOS evidence pack and result-to-claim mapping. Use before paper polishing, revision, compilation, or submission when the user needs a deterministic evidence-boundary report.
Prepare a ResearchOS manuscript as a submission bundle and compile it with the configured native or Docker LaTeX backend. Use when the user needs a real PDF, compile diagnostics, and an auditable summary rather than shell-only LaTeX instructions.
Recover section-level evidence from existing ResearchOS paper notes for one claim or manuscript section. Use when a draft, idea, or review needs grounded evidence without re-reading the entire literature corpus or inventing citations.
Mine local reference research-agent repositories into traceable, transferable ResearchOS pattern cards and a transfer matrix. Use when a user supplies local source locations and wants engineering lessons, not copied implementation or unsupported claims.
Normalize a research topic and user materials into a bounded, evidence-aware scope brief before literature search or ideation. Use when a researcher needs to turn an informal topic into clear questions, exclusions, success criteria, and an upload inventory.
Review an already compiled research manuscript for submission readiness, covering traceable artifacts, citation/claim checks, PDF status, anonymity, and unresolved blockers. Use after paper-compile and before actual venue submission.
Generate the one permitted deterministic taxonomy-overview PDF for a ResearchOS survey. It uses only explicit taxonomy labels and direct paper-ID links from the survey plan, and skips rather than inventing a visual.
LLM guidance for mechanism-aware novelty auditing.