timsmykov
UserAgent-first Evidence Lab research plugins and deterministic onboarding for Codex and Claude Code
Categories
Indexed Skills (29)
paper-lookup
Search 11 academic literature APIs for papers, preprints, citations, and open-access full text, and return results with reproducible provenance. Covers PubMed, PMC (full text), Europe PMC (full-text and preprint search), bioRxiv, medRxiv, arXiv, OpenAlex, Crossref, Semantic Scholar, CORE, Unpaywall. Use when searching for papers, citations, DOI/PMID/arXiv lookups, abstracts, full text, open-access PDFs, preprints, citation graphs, author publications, or any scholarly literature query. Triggers on mentions of any supported database or requests like "find papers on X", "look up this DOI", "who cites this paper", or "get me the PDF".
scientific-critical-thinking
Audit scientific claims, assumptions, causal language, bias, confounding, and evidence quality without drafting a formal referee report. Use for critical appraisal, evidence grading, or teaching claim evaluation; use peer-review for a manuscript review and statistical-analysis for new calculations.
scientific-visualization
Create and audit truthful, accessible, publication-ready scientific figures with Matplotlib, Seaborn, or Plotly. Use for figure design, multi-panel layouts, uncertainty and missing-data displays, color/contrast review, image metadata validation, and journal export planning.
statistical-power
Sample-size and statistical power calculations for planning studies. Use whenever someone asks "how many subjects/samples/replicates do I need", wants an a priori power analysis, a minimum detectable effect (MDE), a power curve, or needs to justify a sample size for a grant, IRB protocol, or pre-registration. Covers closed-form power for t-tests, ANOVA, proportions, correlations, chi-square, and regression, plus simulation-based (Monte Carlo) power for designs with no formula — logistic/Poisson regression, mixed models, cluster-randomized trials, survival, and interactions. Use this skill even when the request only mentions an effect size, alpha, or "80% power" without saying "power analysis" explicitly. For laying out the study (randomization, blocking, factorial/DOE, crossover, sequential designs) use experimental-design; for analyzing data already collected and reporting it use statistical-analysis.
uncertainty-and-units
Track physical units and propagate measurement uncertainty in scientific calculations using pint and uncertainties. Use for unit conversion and dimensional checking, GUM uncertainty budgets, Type A and Type B evaluation, coverage factors and expanded uncertainty, Monte Carlo propagation, significant-figure and plus-minus reporting, error propagation through curve fits, CODATA constants, auditing Python code for stripped units or broken uncertainty propagation, and order-of-magnitude plausibility checks using dimensionless groups (Reynolds, Peclet, Damkohler, Knudsen, Biot, Womersley), characteristic scales such as diffusion time or Debye length, and observed magnitude ranges. Trigger on "is this number physically reasonable", "sanity check these units", "what regime is this flow in", or a result that looks off by orders of magnitude.
literature-review
Plan and conduct reproducible literature reviews with explicit search boundaries, documented screening decisions, structured evidence extraction, quality appraisal, synthesis, and verified citations. Use for systematic, scoping, rapid, or narrative reviews; use paper-lookup for a bounded paper search and writing-skill only after the evidence set is stable.
markdown-mermaid-writing
Create versionable Mermaid diagrams and apply consistent Markdown structure to research artefacts. Use for workflows, timelines, concept maps, architectures, screening flows, and other structural diagrams; use scientific-visualization for data-derived charts and writing-skill for prose drafting.
experimental-design
Design experiments and studies BEFORE data is collected — choosing a design, randomizing, blocking, and laying out treatment combinations so results are interpretable. Use whenever someone is planning a study, asks how to assign subjects/samples to groups, mentions randomization, blocking, stratification, controls, factorial or fractional-factorial designs, design of experiments (DOE), screening many factors, response-surface optimization, crossover or repeated-measures or split-plot designs, cluster/group randomization, Latin squares, plate layouts, batch/run-order effects, replication vs. pseudoreplication, or sequential/adaptive/group-sequential designs. Trigger even for informal phrasings like "how should I set up this experiment", "how do I avoid confounding", "what's the best way to test these 6 factors", or "assign these mice to conditions". For computing the sample size or power once the design is chosen, use statistical-power; for analyzing data already collected, use statistical-analysis.
database-lookup
Query documented public database APIs with explicit endpoints, filters, pagination, and provenance. Use when a scientific, regulatory, financial, or other database-backed fact must be retrieved reproducibly from a named source rather than inferred from general knowledge.
exploratory-data-analysis
Perform bounded, local exploratory analysis of explicitly supported scientific files. Use for redacted CSV/TSV/JSON profiles; optional NumPy, HDF5, FASTA/FASTQ, and basic image metadata inspection; missingness/leakage audits; outlier and transformation sensitivity; and rigorous EDA report scaffolds. Other domain formats are reference-only and unknown formats fail closed.
statistical-analysis
Guided statistical analysis for research data - test selection, assumption checking, effect sizes, power analysis, Bayesian alternatives, and APA-formatted reporting. Use whenever a user wants to compare groups, test a hypothesis, analyze experimental or survey data, check statistical assumptions, compute required sample sizes, or write up results - even if they never name a specific test. Covers t-tests, ANOVA, chi-square, correlation, regression, non-parametric and Bayesian methods. Low-level model implementation requires an explicitly selected and approved library workflow.
markitdown
Convert heterogeneous documents and selected URIs to Markdown with Microsoft MarkItDown for text analysis, search, and LLM/RAG ingestion. Use when a supported PDF, Office, data, archive, audio, or web source must become normalized Markdown; use writing-skill only when the task is to author or revise prose.
peer-review
Prepare evidence-bounded, constructive peer-review drafts and structured manuscript assessments. Use for authorized review of scientific manuscripts, protocols, preprints, or research proposals; reporting-guideline selection; claim–evidence checks; methods, statistics, reproducibility, ethics, figure/table, and citation critique; or revision-response planning.
writing-skill
Create, rewrite, adapt, shorten, expand, audit, and polish evidence-bounded Russian or English text across academic, professional, essay, copywriting, and creative modes. Use when asked to draft or revise text, calibrate voice, turn notes into prose, or diagnose writing quality. Do not use for literature discovery, citation retrieval, peer review, statistical analysis, or document-file manipulation without a writing task.
hypothesis-generation
Formulate evidence-bounded scientific questions, candidate hypotheses, rival explanations, causal or associational claims, discriminating predictions, measurements, and preregistration-ready analysis plans. Use when turning observations or preliminary findings into transparent, testable research plans without treating hypotheses as facts.
evidence-lab-onboarding
Runs the fast Evidence Lab setup when a researcher asks to configure, personalize, install, or choose research capabilities; do not trigger for ordinary research questions after setup.
personal-skill-authoring
Turns a repeated personal research task into a small, portable, tested agent skill. Use when a researcher asks to create, teach, save, automate, improve, or package their own recurring workflow for Codex or Claude Code. Do not use merely to perform a one-off research task.
example-procedure
Reference skill: turns a folder of source documents into a reviewed summary table with an explicit human confirmation step. Loads on requests like 'build a summary table from these documents', 'structure this collection', 'pull these files into one comparable table'. Does NOT load for finding new sources, for checking a finished artefact (use example-checklist), or for questions about the method itself. Ships as a format example — it demonstrates the split between model reasoning and deterministic scripting, not a real research methodology.
task-intake
Convert incoming work into atomic, well-written task cards with accountable ownership, execution and review roles, project context, observable results, acceptance criteria, dependencies, readiness gates, and deterministic priority order. Use when someone asks to create, record, triage, assign, prioritize, reorder, complete, or archive tasks in any project or shared work queue. Do not use for explaining prioritization without changing tasks, product planning from scratch, or inventing work outside the named request or source.
evidence-lab-meeting-capture
Turn an Evidence Lab meeting transcript, recording note, or rough minutes into a source-bounded summary and save the finished page in the canonical Notion meeting register. Use when asked to summarize, capture, document, or file an Evidence Lab meeting, including course, consulting, product, or team discussions. Do not use for ordinary meetings outside Evidence Lab, transcript extraction alone, or task-board intake without a meeting summary.
example-checklist
Reference skill: checks a finished artefact against formal criteria before it goes to a supervisor or a client. Loads on 'review this artefact for gaps', 'what won't survive scrutiny here', 'check this table before I send it'. Does NOT load when the artefact still has to be produced (use example-procedure), when the ask is to rewrite the text rather than check it, or when someone wants a subjective opinion on quality. Ships as a format example: it shows why a plugin bundles several skills instead of shipping one.
life-science-protocols
Add life-science-specific protocol, biological-control, bias, biosafety-boundary, reporting, and acquisition-metadata checks to a study or image workflow. Use when a life-science project is being planned or biological images are analysed; do not provide clinical decisions or experimental safety authorization.
publication-monitoring
Set up or update a repeatable scholarly-publication monitor with explicit sources, queries, date boundaries, stable identifiers, deduplication, and a source-linked digest. Use for recurring alerts or update scans; do not use for a one-time paper lookup or claim exhaustive coverage of inaccessible databases.
qualitative-analysis
Develop and apply a traceable qualitative codebook to interviews, open-ended responses, field notes, or text corpora, preserving coded excerpts, reflexive notes, disagreements, and negative cases. Use for qualitative coding and thematic synthesis; do not substitute generic summarization or tabular statistics.
research-image-analysis
Inspect scientific or research images while preserving source identity, acquisition context, transformations, measurement definitions, exclusions, and interpretation limits. Use for microscopy, experimental photographs, scans, or instrument images; do not use merely to make decorative figures or infer unsupported diagnoses.
systematic-review
Build and run a reproducible systematic-search, deduplication, and screening workflow with explicit eligibility criteria and decision reasons. Use for systematic, scoping, or evidence reviews that require an auditable record flow; do not use for an informal literature overview or fabricate database coverage.
__skill__
REPLACE ME with 2-4 sentences a router can act on. State what the skill produces, then name the concrete phrasings that should load it — 'screen these abstracts', 'build a PRISMA flow', 'check this corpus for missing seminal papers'. Then name the near misses that must NOT load it. Write triggers, not marketing.
data-and-pdf-router
Explain the compatibility bundle that combines document evidence extraction with structured-data analysis. Use only when one complete workflow genuinely needs both PDFs or papers and tables or datasets; document-only and data-only requests should load a focused pack instead.
full-research-cycle-router
Explain the complete Evidence Lab workflow bundle and confirm that the researcher wants both research design and literature or publication support. Use only when the researcher explicitly asks for the full research cycle; focused planning, review, or writing requests should load the narrower pack directly.
Bio shown is the top-scored skill's repo description as a fallback — real GitHub bios land in a future update.