Full systematic literature review (PRISMA 2020 core) with discovery, screening, evidence extraction, synthesis, confidence grading, and adversarial stress-testing across papers. Prefer peer-reviewed published versions over preprints and use this skill either directly or as the coordinated literature stage inside a larger research workflow. Use when asked to run or audit a literature review, evidence synthesis, state-of-the-art survey with explicit methods, citation integrity checks, or confidence-rated conclusions from research publications.
domain before substantive review work. Stop and ask if missing.technical_exposition is not provided, explicitly ask the user to choose between standard vs detailed_math. Apply defaults for missing non-domain inputs and log assumptions.research-zotero, or a user corpus.scripts/init_review_pack.py.research-zotero first or consume existing ./zotero/ artifacts.scripts/prisma_flow_md.py and insert it into <topic>.review.md.scripts/validate_review_pack.py before returning output.domain: mandatory. Do not continue without this value.research_question: default to "What does current evidence show about <topic> in <domain>?"inclusion_criteria: default to peer-reviewed primary studies (published) relevant to topic and domain; use high-quality preprints only when no published version exists or when recency is essential, and label them.exclusion_criteria: default to off-topic studies, non-substantive summaries, and sources without sufficient methodological detail.date_range: default to the last 10 years through today.study_types: default to experimental, observational, benchmarking, and systematic-review studies when relevant.language: default to English.population/context: default to the broad population/context implied by topic and domain.outcomes: default to efficacy/performance, robustness, safety, and transferability outcomes where applicable.quality_threshold: default to retain studies with at least moderate methodological quality and transparent reporting.technical_exposition: default to standard. If set to detailed_math, the <topic>.review.md must include formal definitions/notation, key equations/objectives, and (when relevant) theorem statements or proof sketches in the Synthesis section.zotero_library_type: optional. user or group when Zotero should be used as a source.zotero_library_id: optional. Zotero user ID or group ID.zotero_collection_key: optional. Restrict Zotero sync to a collection.zotero_query: optional. Zotero free-text query for item discovery.zotero_tags: optional. Restrict Zotero items by tag.zotero_access_mode: optional. api-key, oauth-key, or mcp.domain is missing.research-zotero artifact or Zotero API/MCP path.Primary file:
<topic>.review.md
Supporting files:
<topic>.protocol.md
<topic>.search-log.md
<topic>.screening-log.md
<topic>.evidence-table.md
Optional support files when Zotero is used:
<topic>.zotero-items.json
<topic>.zotero-sync.md
Required sections in <topic>.review.md:
Protocol
Search Strategy
Screening Decisions
Evidence Table
Synthesis
Adversarial Stress Test
Limitations
Confidence Assessment
PRISMA flow accounting
<topic> to lowercase hyphen-case for file names.references/protocol-template.md.references/search-strategy-template.md.research-zotero or consume its artifacts and decide whether Zotero is:
./zotero/zotero-items.json already exists, prefer consuming it over re-syncing.research-zotero should prefer it for interactive library inspection.research-zotero should export items and log the sync before this skill consumes them.api2.openreview.net) for reproducible queries (e.g., notes/search + notes?id=...), and log venue group(s), query terms/fields (title/abstract), and whether you filtered to accepted papers.venue / venueid cues; avoid treating Rejected_Submission / Withdrawn_Submission / “Submitted to …” as canonical). Keep arXiv as an access copy when helpful.references/screening-template.md.duplicate-publication with a note like "preprint superseded by published version".references/evidence-table-template.md.references/report-template.md.technical_exposition=detailed_math, make the Synthesis section math-forward: define the core objects precisely, write the primary learning objectives/constraints, and summarize theoretical results using correct formal statements (without over-quoting).references/adversarial-literature-checklist.md.references/domain-adapters.md.scripts/prisma_flow_md.py.scripts/validate_review_pack.py.scripts/init_review_pack.py: create deterministic Markdown scaffolds for protocol/search/screening/evidence/report.scripts/prisma_flow_md.py: parse standardized screening counts and emit PRISMA flow accounting Markdown.scripts/validate_review_pack.py: validate required sections, mandatory fields, and PRISMA count consistency.references/protocol-template.mdreferences/search-strategy-template.mdreferences/screening-template.mdreferences/evidence-table-template.mdreferences/prisma-core-checklist.mdreferences/adversarial-literature-checklist.mdreferences/domain-adapters.mdreferences/report-template.md../research-zotero/references/zotero-artifact-contract.md