| name | plan-a-literature-review |
| category | school |
| description | Plan a transparent academic literature review with a focused question, reproducible search strategy, screening rules, evidence table, synthesis method, and citation boundaries. Use when a student or researcher must survey scholarship systematically without having the agent write graded analysis or invent sources. |
plan-a-literature-review
Build a review process another reader can inspect. Make the question, search, inclusion, exclusion, coding,
and synthesis decisions explicit before drawing conclusions.
Inputs
- Obtain the assignment or protocol, topic, discipline, date range, databases, deadline, and course AI policy.
- Name the review type, reviewer, supervisor, citation style, and accessible full-text sources.
- Keep the student responsible for reading, interpretation, and graded writing.
Procedure
- Turn the topic into a focused review question with population or context, concept, outcome, and boundaries where useful.
- Define the purpose and review type: narrative, scoping, systematic, rapid, or another approved method.
- Build concepts, synonyms, controlled vocabulary, exclusions, date and language limits, and a pilot search string.
- Choose databases and supplementary sources appropriate to the discipline. Record access dates and coverage limits.
- Define inclusion and exclusion criteria before screening results.
- Pilot the search and test whether known relevant work appears. Revise transparently.
- Save each exact query, database, filters, date, and result count. Deduplicate records without losing source ids.
- Screen title and abstract, then full text, recording reasons for exclusion at the required level.
- Create an evidence table with citation, method, sample, context, finding, limitation, and relevance.
- Assess source quality using an appropriate rubric and distinguish study limitations from disagreement.
- Plan synthesis by themes, methods, chronology, theory, or evidence strength, preserving counterevidence and gaps.
- Verify every citation from the source before drafting in the student's own words.
Boundaries
Do not invent citations, summaries, page numbers, findings, or access. Never bypass paywalls or submit
agent-written graded analysis as the student's work. Follow research ethics and course rules.
Done
- A review protocol states question, type, databases, search strings, dates, and screening rules
- A deduplicated record log preserves source ids, decisions, and exclusion reasons
- An evidence table records verified methods, findings, limitations, relevance, and citation details
- A synthesis outline distinguishes evidence patterns, counterevidence, quality, gaps, and student-authored interpretation
Then use research-paper-workshop for the student's drafting process.
