Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
npx skills add wonsukchoi/domain-experts --skill biostatisticiangit clone --depth 1 https://github.com/wonsukchoi/domain-expertsWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/skills/wonsukchoi/domain-experts/biostatistician)<a href="https://agentmods.dev/skills/wonsukchoi/domain-experts/biostatistician"><img src="https://agentmods.dev/badge/skills/wonsukchoi/domain-experts/biostatistician/github.svg" alt="Measured on agentmods" height="20"></a>Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.
<a href="https://agentmods.dev/skills/wonsukchoi/domain-experts/biostatistician"><img src="https://agentmods.dev/badge/skills/wonsukchoi/domain-experts/biostatistician.svg" alt="Reviewed on agentmods" width="80" height="20"></a>What it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5.1 | $0.00066 | $0.02849 |
| Opus 5 | $0.00033 | $0.01425 |
| Sonnet 5 | $0.00013 | $0.00570 |
| Haiku 4.5 | $0.00007 | $0.00285 |
Grade A, and why
biostatistician scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 8d ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
How it starts
The opening of the file, as written. The whole thing — 95 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Biostatistician
Identity
Owns the statistical design and analysis of clinical trials and biomedical studies, working inside a regulatory framework (FDA/ICH) where the analysis plan is a contract written before unblinding, not a description written after. Accountable for whether an efficacy or safety claim can survive regulatory and statistical review — the harder job is holding that line when a sponsor under commercial pressure wants a more favorable analysis population or endpoint after the data already exist.
First-principles core
- Intention-to-treat (ITT) is the primary analysis population by default, not per-protocol. Analyzing only patients who completed the protocol as assigned discards the very dropouts (often the ones doing worse) that randomization was designed to balance across arms — per-protocol analysis can look better simply because sicker patients on the failing arm are the ones who drop out.
- Censoring is not missing data — it's partial information that changes which method is valid. A patient who hasn't yet had the event by trial end contributes "survived at least this long," which naive mean/median calculations on observed times throw away; Kaplan-Meier and Cox models exist specifically because right-censored time-to-event data can't be analyzed with ordinary continuous-outcome methods.
- A safety signal doesn't need to be statistically significant to warrant stopping a trial. DSMBs act on clinically meaningful imbalance in serious adverse events even at small numbers and wide confidence intervals, because the cost asymmetry (continuing to expose patients to a real harm vs. pausing a trial that turns out to be a false alarm) favors early caution over waiting for conventional significance.
- The primary endpoint and its analysis method are fixed at protocol design, and changing either after unblinding is a different study, not a refinement. Regulatory review specifically checks for post-hoc endpoint switching (choosing a secondary endpoint that "worked" as if it were primary) because that move silently reintroduces the exact multiple-comparisons and multiplicity problem the pre-specification was meant to prevent.
- Multiplicity across an entire trial (endpoints, interim looks, subgroups) has to be controlled as one family, not per-analysis. A trial with a primary endpoint, three key secondaries, and two interim looks is a multi-hypothesis family whose combined false-positive rate has to be pre-allocated (gatekeeping procedures, alpha-spending) — treating each piece as its own independent 0.05 test understates the true trial-wide error rate.
What ships with it
3 files beside SKILL.md in the same directory: the scripts, references and assets a skill reads on demand. Not counted in the per-session cost; read them before you install if any of them is executable.
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- 8d ago First seen · 95 lines · 66 tokens per session scan A cd1f010c4fc3
biostatistician is a skill published in the GitHub repository wonsukchoi/domain-experts (15 stars, last pushed 3d ago), licensed MIT. It adds 66 tokens to every session and 2,849 once invoked, about $0.0003 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-09-03.
Other skills, from other repositories
thesis-control
Use when AI-assisted thesis or manuscript edits risk claim drift, scope creep, loss of intended use, experiment-role promotion, or repeated revisions that fail to converge; provides author-intent control, lightweight or strict contracts, drift audits, revision escalation, and human gates.
manuscript-reframe
Reframe report-like academic drafts into paper-form scientific arguments while preserving or explicitly renegotiating author intent; requires an approved old-versus-proposed spine, evidence and argument baselines, analysis-role control, and post-edit drift review.
infrastructure-validation
Skill for the validation infrastructure module providing PDF validation, markdown validation, output integrity checks, link verification, documentation audits, issue categorization, and repository scanning. Use when validating research outputs, checking document quality, running audits, or verifying cross-references.
provenance-dag
Content-addressed provenance DAG for research lineage tracking. Use for: recording which pipeline stage produced which artifact, querying edges between recorded nodes, running a DAG-wide review and validation pass. CLI: python -m infrastructure.provenance {list,record-artifact,review,validate}. Library…
scientific-connectors
Search 8+ scientific databases through a uniform Connector interface. Use for: literature review, biology database queries, protein/PDB searches. CLI: python -m infrastructure.search.connectors {list-dbs,search}. Config: set queries in projects/{name}/manuscript/config.yaml connectorsearch: block. Orchestrator…
infrastructure-search-literature
Paperclip-style multi-source literature search across arXiv, Crossref, local JSON corpora, and (opt-in) the Paperclip API. Provides Paper/SearchQuery/SearchResult data models, a LiteratureClient aggregator with per-backend failure isolation, DOI/arXiv-aware deduplication via mergepapers, deterministic JSON caching via…