coleoguy
UserClaude Skill to collect data from literature
Categories
Indexed Skills (9)
traittrawler
Trait-and-clade-agnostic scientific literature mining pipeline. Given any trait and any taxonomic scope, TraitTrawler bootstraps from existing curated data (optional), learns how the trait is reported in the literature, proposes an output schema AND candidate validation hooks for the user to approve, and then autonomously searches, fetches, and extracts structured records into a verified CSV plus a full per-row audit ledger. Grounding is a protocol invariant — every row ties to a SHA256-hashed PDF, a page number, and a verbatim quote that a deterministic validator has already confirmed appears in that PDF. Use when the user mentions trait extraction, literature mining, database building, phenotype harvesting, systematic review data collection, or anywhere else they want structured data from a corpus of papers.
grant-section-drafter
Draft one section of a grant application, manuscript, cover letter, addendum, or any extended prose that must read in the researcher's own voice. Use when Heath asks for a first-pass draft of a Significance section, a Specific Aim, an Approach narrative, a rebuttal, a cover letter paragraph, or any piece of scientific prose longer than ~200 words that will be reviewed line-by-line or submitted to a funding agency, journal, or search committee. Do NOT use for short chat replies, tool-output summaries, or bullet-point notes.
hypothesis-pipeline-rubric
Evaluate, propose, or critique a scientific hypothesis or testable claim using the lab's type-aware gating pipeline. Use when a chat message or scheduled job produces a directional, mechanistic, comparative, methodological, or synthesis claim that should be checked for sign coherence, mechanism articulation, falsifiability, and novelty relative to prior work. Also use when explaining why a hypothesis was blocked or borderline, or when coaching a new claim to meet the pipeline bar before submission to the formal pass.
karyotype-databases
Use when working with karyotype, chromosome number, sex-system data, or any of the lab's curated species databases (Coleoptera, Diptera, Amphibia, Mammalia, Polyneoptera, Drosophila karyotypes; Tree of Sex; CURES; Epistasis; Tau). Covers how to resolve resource keys via require_data_resource, what each database contains, and safe read/write patterns.
paper-reviewer
Conduct a formal peer review of a scientific paper Heath did NOT author — for a journal he is reviewing for, or as a pre-submission read for a collaborator. Six-phase workflow with a Coordinator (segments paper + routes), 6 parallel specialist subagents (Methods Auditor, Logic Checker, Presentation, Journal Fit, Adversarial Reader, Citation Verifier), a Synthesizer, and a Refiner that applies the structural rubric (blocking vs advisory tags, 3:1 minor:major cap, action-oriented minor comments). Output is a markdown review tuned to Heath's voice via the bundled voice.md. Bundle reference content lives at `Shared drives/Blackmon Lab/ Projects/paper-reviewer/` (agents/, references/, checklists/, voice.md). TRIGGER on: "review this paper", "peer review", "referee for <journal>", "what do I think of this manuscript", a directory of review materials, mention of revision/rebuttal/response-to-reviewers. Distinct from `pre_submission_review` (which is for Heath's OWN drafts heading to submission, not papers he is revi
pre-submission-review
Pre-submission review of a manuscript Heath authored or co-authored, before it goes to a journal or funder. The output is ALWAYS a Google Doc copy of the manuscript with visible inline review markup (red strikethrough for old text + red insertion for new text — both visible side-by-side) plus margin comments. NEVER a journal-style prose review written into a fresh blank doc. Same multi-agent architecture as paper-reviewer (Coordinator → 6 parallel specialists → Synthesizer → Refiner) with three critical inversions: stance is INTERNAL ADVISOR ("find issues so the author can fix them, not so a reviewer can score them"), voice is HEATH'S MANUSCRIPT voice (NOT peer-review voice), output is a marked-up copy via mark_changes_in_google_doc + insert_comment_in_google_doc on a doc made by copy_google_doc. TRIGGER on any phrasing where Heath wants HIS OWN manuscript reviewed before submission. All of these route here: "review my paper before I submit" "pre-submission review" "review the <topic> manuscript" / "review th
r-comparative-phylogenetics
Use when writing R code for comparative phylogenetic analysis (BiSSE/MuSSE, ancestral state reconstruction, BAMM, diversitree, sex-chromosome turnover, dysploidy rate analysis). Covers the standard library set, code templates, the data-resource preflight rule, working-directory conventions, and sandbox restrictions.
voice-matching
Write extended prose (>~150 words) that reads as the researcher's own writing, not as AI-generated text. Use when producing grant sections, cover letters, addenda, rebuttals, manuscript drafts, lab-website updates, or any output that will be read by a program officer, journal reviewer, or hiring committee as the researcher's first-person work. The voice index (agent/voice_index.py) provides retrievable exemplars of the researcher's published prose ranked by topic similarity; this skill describes how to use them and what patterns to actively avoid.
wiki-authoring
Use when authoring or editing pages under /knowledge/ on the lab website (paper pages, topic pages, repo pages). Covers required frontmatter fields, slug rules, the h1-must-match-title invariant, finding anchors, git commit conventions, the wiki_janitor audit, and the preservation rule for category and papers_supporting fields.
Bio shown is the top-scored skill's repo description as a fallback — real GitHub bios land in a future update.