neelshah4
UserClaude Code plugin: multi-agent critical-care consult. Decision support for licensed clinicians.
Categories
Indexed Skills (8)
fabrication-audit
Silent post-draft verification pass catching unverified specifics before they leave the response. Triggers on ANY substantive output containing factual claims about the world: numbers, dates, addresses, jurisdictions (which office/agency covers a place), distances, drive times, government fees, processing times, version numbers, current officeholders, drug doses, prices, statistics, sports stats/cap hits, ETF/stock specifics, salaries, antimicrobial susceptibility/resistance rates — any "the X is Y" / "X serves Y" pattern where Y is a specific retrievable value. Embedded sub-claims in longer responses get same treatment as direct questions. Triggers on nearly every substantive turn except greetings, pure code, pure math derivations, fictional creative writing. Does NOT trigger on academic citations (PMIDs/DOIs/journal refs) — citation-verification handles those exclusively. Runs silently, fixes draft before sending. Err very heavily on triggering. When in doubt: TRIGGER.
icu-clinical-consult
Multi-agent adversarial clinical consult for pediatric (and adult) critical care questions. Triggers on ANY clinical question involving a patient scenario, physiologic mechanism, ventilator management, hemodynamics, ECMO, shock, sepsis, neurocritical care, renal, cardiac perioperative care, antimicrobial selection, or any bedside ICU question. Also triggers for mechanism-level physiology questions (e.g., "how does X affect Y") even without a specific patient. Trigger phrases include: "clinical question," "I have a patient," "what would you do with," "management of," "mechanism of," "physiology of," any drug/vent/hemodynamic question, antimicrobial spectrum/coverage question, differential diagnosis, "what's the evidence for," or any scenario that a PICU attending would reason through. Use this skill even for seemingly simple clinical questions — err on the side of triggering. Do NOT use for grant writing, manuscript editing, literature reviews, data analysis, or non-clinical academic tasks.
fabrication-audit
Silent post-draft verification pass catching unverified specifics before they leave the response. Triggers on ANY substantive output containing factual claims about the world: numbers, dates, addresses, jurisdictions (which office/agency covers a place), distances, drive times, government fees, processing times, version numbers, current officeholders, drug doses, prices, statistics, sports stats/cap hits, ETF/stock specifics, salaries, antimicrobial susceptibility/resistance rates — any "the X is Y" / "X serves Y" pattern where Y is a specific retrievable value. Embedded sub-claims in longer responses get same treatment as direct questions. Triggers on nearly every substantive turn except greetings, pure code, pure math derivations, fictional creative writing. Does NOT trigger on academic citations (PMIDs/DOIs/journal refs) — citation-verification handles those exclusively. Runs silently, fixes draft before sending. Err very heavily on triggering. When in doubt: TRIGGER.
citation-verification
Verify every academic citation and reference in real time. Triggers whenever Claude cites a paper, recommends a reference, quotes evidence, or produces output containing author-year citations, DOIs, PMIDs, or journal references — manuscripts, grants, abstracts, literature reviews, clinical consults with cited evidence, rebuttal letters, teaching materials, any response where a specific study is named. Scope is academic citations ONLY — for non-academic verifiable specifics (jurisdictions, fees, addresses, free-standing drug doses, distances, prices, sports/finance specifics), fabrication-audit handles those. The two skills compose. Runs as silent verification pass — no separate output unless verification fails or user requests audit. Never fabricate a citation. If unverifiable, mark [CITATION NEEDED]. Err heavily on triggering — any time a paper is mentioned by name, this skill should be active.
fabrication-audit
Silent post-draft verification pass catching unverified specifics before they leave the response. Triggers on ANY substantive output containing factual claims about the world: numbers, dates, addresses, jurisdictions (which office/agency covers a place), distances, drive times, government fees, processing times, version numbers, current officeholders, drug doses, prices, statistics, sports stats/cap hits, ETF/stock specifics, salaries, antimicrobial susceptibility/resistance rates — any "the X is Y" / "X serves Y" pattern where Y is a specific retrievable value. Embedded sub-claims in longer responses get same treatment as direct questions. Triggers on nearly every substantive turn except greetings, pure code, pure math derivations, fictional creative writing. Does NOT trigger on academic citations (PMIDs/DOIs/journal refs) — citation-verification handles those exclusively. Runs silently, fixes draft before sending. Err very heavily on triggering. When in doubt: TRIGGER.
prompt-optimizer
Use when sharpening a user's request before producing content — fires on nearly every substantive turn that carries a task, even when unasked. Trigger on: a prompt past one sentence with a task; any content-production ask (writing, analysis, code, slides, grants, manuscripts, emails, reasoning); an attachment-led ask (file plus instruction); a pasted plan or persona to execute or critique; a skill/hook/agent-engineering ask; a short but substantive ask; or a cue (optimize my prompt, help me draft/analyze/build/review). When ambiguous, fire. Skip categories OVERRIDE the trigger: greetings; bare acks/continuations; single-fact lookups with no synthesis; bare yes/no; bare definitions; status/verification checks; deterministic file/tool commands; in-prompt continuations or edits (resume, continue, check it, make the Nth shorter), though a continuation adding new work fires; or a block longer than the answer. Emits a 5-element prompt block on the first response per task.
fabrication-audit
Silent post-draft verification pass catching unverified specifics before they leave the response. Triggers on ANY substantive output containing factual claims about the world: numbers, dates, addresses, jurisdictions (which office/agency covers a place), distances, drive times, government fees, processing times, version numbers, current officeholders, drug doses, prices, statistics, sports stats/cap hits, ETF/stock specifics, salaries, antimicrobial susceptibility/resistance rates — any "the X is Y" / "X serves Y" pattern where Y is a specific retrievable value. Embedded sub-claims in longer responses get same treatment as direct questions. Triggers on nearly every substantive turn except greetings, pure code, pure math derivations, fictional creative writing. Does NOT trigger on academic citations (PMIDs/DOIs/journal refs) — citation-verification handles those exclusively. Runs silently, fixes draft before sending. Err very heavily on triggering. When in doubt: TRIGGER.
grant-reviewer
Multi-agent grant reviewer producing scored critiques and copy/paste-ready edits to maximize funding probability. Triggers on: 'review my grant,' 'grant review,' 'score my aims,' 'NIH review,' 'study section,' 'pink sheet,' 'edit my R01/R21/K08/K23/R03/U01,' 'strengthen my aims,' 'resubmission,' 'mock study section,' or any request to evaluate a research grant. Mechanism-adaptive: R-series use the NIH Simplified Review Framework (Factor 1/2/3); K-series score the five K criteria (Candidate, Career Development Plan, Research Plan, Mentors, Environment). Also DoD CDMRP, PCORI, AHRQ, foundations. Nine agents: Storyboard, Rigor, Stats, Payoff Aim, Compliance, Positioning, conditionally Inclusion, Budget, and Career Development (K). Phase 2.5 Cross-Check, Phase 3.5 Edit Verification, Phase 4 two-phase Study Section, Phase 4.5 Reviewer-2 Steelman; depth modes pink/red/gold. Outputs: scored edit bank, study-section transcript, .docx pink sheet. Trigger even for 'look at my grant.'
Bio shown is the top-scored skill's repo description as a fallback — real GitHub bios land in a future update.