From grimoire
Plans and writes systematic reviews of scientific literature following PRISMA guidelines — from protocol registration through search, screening, risk-of-bias assessment, and evidence synthesis.
How this skill is triggered — by the user, by Claude, or both
Slash command
/grimoire:write-systematic-reviewThe summary Claude sees in its skill listing — used to decide when to auto-load this skill
Conduct and report a systematic review following PRISMA 2020 standards — with pre-registered protocol, comprehensive search, transparent screening, risk-of-bias assessment, and synthesis of evidence quality.
Conduct and report a systematic review following PRISMA 2020 standards — with pre-registered protocol, comprehensive search, transparent screening, risk-of-bias assessment, and synthesis of evidence quality.
Adopted by: Cochrane Collaboration (the gold standard for systematic reviews), WHO, NIH systematic evidence review programs, NICE (UK clinical guidelines), FDA Evidence-Based Medicine framework, and virtually all clinical guideline bodies require systematic reviews as the primary evidence base for recommendations. Impact: Ioannidis (2005) "Why Most Published Research Findings Are False" demonstrated that individual studies frequently yield false positives — systematic reviews, by pooling and assessing evidence quality, are the most reliable way to estimate true effect sizes. The PRISMA statement is endorsed by >700 journals (EQUATOR network); systematic reviews using PRISMA have significantly higher citation rates and lower retraction rates than narrative reviews.
Pre-registration is mandatory for credible systematic reviews:
Structure the question:
Example: In adults with Type 2 diabetes (P), does GLP-1 agonist therapy (I) compared to SGLT-2 inhibitors (C) reduce cardiovascular mortality (O)?
Search must be comprehensive and reproducible:
Document the exact search string for each database (required for PRISMA flow diagram).
Two-phase screening:
Use systematic review software (Covidence, Rayyan, DistillerSR) to manage screening and track decisions.
Calculate inter-rater reliability (Cohen's kappa ≥ 0.7 = acceptable agreement).
Create a data extraction form before extraction begins:
Extract with two independent reviewers for key outcome data; verify against source.
Every included study must be assessed for methodological quality:
Summarize at the study level and domain level. High-risk-of-bias studies downgrade the certainty of evidence.
Narrative synthesis: when pooling is not appropriate (heterogeneous methods/populations); describe patterns, consistencies, and gaps
Meta-analysis: when ≥2 studies report the same outcome with compatible methods; use random-effects model (DerSimonian-Laird) as default for clinical heterogeneity; report I² for heterogeneity assessment
GRADE assessment: rate evidence certainty for each outcome:
npx claudepluginhub jeffreytse/grimoire --plugin grimoireConducts PRISMA-standard systematic reviews with protocol development, PICOTS framing, exhaustive searches, screening, and Risk of Bias assessments using Cochrane RoB 2.0 or ROBINS-I.
Guides systematic, scoping, and narrative literature reviews using PRISMA/PRISMA-ScR protocols, Boolean/MeSH search strategies, databases (PubMed, Scopus, Web of Science, Embase), screening, extraction, synthesis, and reporting.
Conducts systematic, scoping, narrative, or meta-analysis literature reviews across academic, biomedical, and technical domains. Uses PICO and structured search protocols.