Skip to content

feat: Add all 17 missing reporting checklists (P1+P2+P3) - #3

Merged
sajor2000 merged 2 commits into
mainfrom
feat/p1-checklist-gaps
Apr 29, 2026
Merged

sajor2000 merged 2 commits into
mainfrom
feat/p1-checklist-gaps

Conversation

@sajor2000

@sajor2000 sajor2000 commented Apr 29, 2026 •

Copy link
Copy Markdown
Owner

Summary

Closes #2. Adds all 17 checklist files that were referenced in the routing map but had no backing content. Plugin goes from 18 to 35 reporting checklists.

P1 — Routing map references (4 files)

Checklist Items Use case
STARD-AI 10 AI-assisted diagnostic accuracy
TARGET 9 Target trial emulation
STROBE-MR 8 Mendelian randomization
STREGA 10 Genetic association (GWAS)

P2 — Base/extension parity (4 files)

Checklist Items Use case
TRIPOD 22 Base prediction model (non-AI)
PRISMA-DTA 27 Diagnostic accuracy SRs
PRISMA-NMA 28 Network meta-analyses
PRISMA-IPD 27 Individual patient data SRs

P3 — Routing hints references (9 files)

Checklist Use case
SQUIRE 2.0 Quality improvement
GRAMMS Mixed-methods
STaRT-RWE Real-world evidence (regulatory)
PRISMA-ScR Scoping reviews
CONSORT-Cluster Cluster RCTs
CONSORT-Adaptive (ACE) Adaptive designs
CONSORT N-of-1 (CENT) N-of-1 trials
QUADAS-2 SR quality assessment
CHARMS SR of prediction models

Test plan

  • bun test tests/frontmatter.test.ts — 272/272 pass
  • bun run release:validate — 55 agents, 40 skills
  • 35 checklist files in ce-code-review/references/
  • Every checklist in routing map and routing hints has a file

🤖 Generated with Claude Code

sajor2000 and others added 2 commits April 29, 2026 13:15
Closes all P1 items from issue #2. Plugin now has 22 reporting
checklist files backing the code review skill.

STARD-AI (10 items): AI-assisted diagnostic accuracy studies.
  Covers AI system description, version/threshold, intended clinical
  role, reference standard independence, data partitioning, subgroup
  specs, failure analysis.
  Ref: Sounderajah et al. Nat Med 2020. PMID: 32514173.

TARGET (9 items): Target trial emulation studies.
  Covers eligibility-treatment-assignment-outcome-time zero-follow up
  protocol mapping, the structured two-column table, and emulation
  assumptions/limitations.
  Ref: Hansford et al. BMJ 2025;388:e081680.

STROBE-MR (8 items): Mendelian randomization studies.
  Covers IV assumptions, data sources (one-sample vs two-sample),
  instrument selection (LD clumping, F-statistic), MR methods, and
  sensitivity analyses (MR-Egger, MR-PRESSO, colocalization).
  Ref: Skrivankova et al. JAMA 2021;326(16):1614. PMID: 34698778.

STREGA (10 items): Genetic association studies (GWAS, candidate gene).
  Covers genotyping platform, QC filters, imputation, population
  stratification, multiple testing, genetic models, winner's curse.
  Ref: Little et al. Genet Epidemiol 2009;33(7):581. PMID: 19278015.

Update plugin README checklist count 16 -> 22.

Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com>
…ssue #2

Plugin now has 35 reporting checklist files — every checklist referenced
in the routing map and routing hints has a backing file.

P2 (4 files):
- tripod-checklist.md: base TRIPOD for prediction models (Moons et al. 2015)
- prisma-dta-checklist.md: PRISMA-DTA for diagnostic accuracy SRs (McInnes et al. 2018)
- prisma-nma-checklist.md: PRISMA-NMA for network meta-analyses (Hutton et al. 2015)
- prisma-ipd-checklist.md: PRISMA-IPD for individual patient data SRs (Stewart et al. 2015)

P3 (9 files):
- squire-checklist.md: SQUIRE 2.0 for quality improvement (Ogrinc et al. 2016)
- gramms-checklist.md: GRAMMS for mixed-methods (O'Cathain et al. 2008)
- start-rwe-checklist.md: STaRT-RWE for regulatory real-world evidence (Wang et al. 2021)
- prisma-scr-checklist.md: PRISMA-ScR for scoping reviews (Tricco et al. 2018)
- consort-cluster-checklist.md: CONSORT cluster RCTs (Campbell et al. 2012)
- consort-adaptive-checklist.md: ACE for adaptive designs (Dimairo et al. 2020)
- consort-nof1-checklist.md: CONSORT N-of-1 / CENT (Vohra et al. 2015)
- quadas-2-checklist.md: QUADAS-2 for SR quality assessment (Whiting et al. 2011)
- charms-checklist.md: CHARMS for SR of prediction models (Moons et al. 2014)

Closes #2.

Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com>
@sajor2000 sajor2000 changed the title feat: Add 4 P1 reporting checklists (STARD-AI, TARGET, STROBE-MR, STREGA) feat: Add all 17 missing reporting checklists (P1+P2+P3) Apr 29, 2026
@sajor2000
sajor2000 merged commit e0c0dcd into main Apr 29, 2026
2 checks passed
@sajor2000
sajor2000 deleted the feat/p1-checklist-gaps branch April 29, 2026 20:41
sajor2000 added a commit that referenced this pull request Oct 21, 2026
MedSci audit gap #3 docs-only fallback (#142): the packaged classroom
installer stays out of scope for now, but the audience that avoids
terminals gets an explicit route - once Claude Code runs on any surface
(desktop app or terminal), the marketplace install, method-setup, and
auto-update are all in-session. Per code.claude.com/docs/en/discover-
plugins, plugin install works on any Claude Code surface.

Closes #142

Co-authored-by: factory-droid[bot] <138933559+factory-droid[bot]@users.noreply.github.com>
sajor2000 added a commit that referenced this pull request Oct 21, 2026
Gaps #3, #5, #7, #10 closed (PRs #150, #151, #153, #154); #1, #6, #8
tracked as issues #141, #144, #146 with scoping; #9 wontfix as #147.

Co-authored-by: factory-droid[bot] <138933559+factory-droid[bot]@users.noreply.github.com>
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

feat: Add STROBE extension checklists (RECORD, RECORD-PE, STROBE-MR, STREGA)

1 participant