literature-review

Name: literature-review
Author: scdenney

34stars0forks

How to install literature-review?

scdenney/open-science-skills/literature-review

$npx -y skills add scdenney/open-science-skills --skill literature-review

Installs into the current project.

Summary

Build or audit a literature review. Use for evidence maps, gap analysis, contribution checks, source verification, and synthesis planning.

Files · 1

View on GitHub

SKILL.md

1	# Literature Review Evidence Mapper
2
3	## Heritage and scope
4
5	This is an original Open Science Skills workflow for experimental and computational social science. It remixes high-level ideas from Cheng-I Wu's Academic Research Skills for Claude Code (CC BY-NC 4.0), especially evidence mapping, source verification, and mode separation between narrative literature review and formal systematic review. It is not a full ARS pipeline and should not copy ARS prose.
6
7	## Instructions
8
9	### 1. Classify the review task
10
11	Decide what the user needs:
12
13	- Narrative/theory review: organize concepts, mechanisms, and debates for an introduction.
14	- Design precedent review: identify prior treatments, measures, samples, estimands, or analysis strategies.
15	- Contribution audit: test whether the claimed gap survives contact with the closest prior work.
16	- Evidence map: summarize what each study establishes, where it applies, and what remains unresolved.
17	- Systematic-review escalation: when the user needs exhaustive search, screening, risk-of-bias, and PRISMA reporting.
18
19	Default to a narrative/evidence-map review unless the user explicitly asks for a systematic review, meta-analysis, or PRISMA-compliant output.
20
21	If the user does want a systematic review, be explicit about what this skill can and cannot do: steps 2-7 below produce the protocol scaffold -- question, boundaries, search strings, inclusion/exclusion criteria, and the evidence-map structure -- but exhaustive multi-database screening, dual-coder risk-of-bias assessment, and PRISMA flow accounting require dedicated tooling and human coders. Deliver the protocol scaffold, recommend registering it (PROSPERO or OSF), and point the user at screening tools (e.g., Covidence, ASReview) for the systematic phase.
22
23	### 2. Define the question and boundaries
24
25	Before summarizing papers, specify:
26
27	- Research question or review question.
28	- Population, setting, outcome, treatment/exposure, and mechanism scope.
29	- Disciplines and literatures that must be included.
30	- Time window and language restrictions, if any.
31	- Inclusion/exclusion logic for sources.
32	- What counts as "closest prior work."
33
34	If the user only gives a broad topic, first produce a short scoping memo with 2-4 possible review boundaries rather than writing a generic review.
35
36	### 3. Build the source base
37
38	Use the user's supplied sources first. Then identify obvious missing source classes:
39
40	- Seminal theoretical anchors.
41	- Most recent directly related empirical work.
42	- Meta-analyses, registered reports, study registries, or working-paper series.
43	- Methods papers that justify the research design.
44	- Null findings, failed replications, or unpublished registered studies when visible.
45	- Adjacent literatures that use different vocabulary for the same construct.
46
47	Run `citation-check` when the source list is large, messy, DOI-heavy, or likely to contain stale working papers.
48
49	### 4. Extract evidence, not summaries
50
51	For each important source, record:
52
53	- Claim actually supported: one sentence, no inflation.
54	- Design and identification: sample, setting, treatment/exposure, comparison, outcome, estimand, and key limitations.
55	- Observable implication: what the source lets a reader expect in the user's setting.
56	- Boundary condition: where the finding may fail.
57	- Use in the user's paper: background, theory, design precedent, measurement precedent, competing explanation, or gap support.
58
59	Do not produce chronological "Author A says X, Author B says Y" prose unless chronology is theoretically important.
60
61	### 5. Cluster the literature
62
63	Organize sources into 3-6 clusters. Prefer conceptual or mechanism clusters over method-only clusters:
64
65	- Mechanism families.
66	- Competing theories.
67	- Measurement traditions.
68	- Empirical settings or populations.
69	- Identification strategies.
70	- Evidence-quality tiers.
71
72	For each cluster, state what is settled, what is contested, and what would change the interpretation.
73
74	### 6. Test the claimed gap
75
76	Write a gap verdict:
77
78	- Holds: no close prior work answers the same question with the same population/mechanism/design.
79	- Partly holds: prior work answers part of it; the contribution must be narrowed.
80	- Does not hold: the claimed contribution is already established; reframe as replication, extension, boundary test, or synthesis.
81	- Cannot assess: missing sources or inaccessible source text prevent judgment.
82
83	When the gap is weak, propose a better contribution frame rather than only criticizing it.
84
85	### 7. Compose with sibling skills
86
87	- Use `narrative-building` after the evidence map exists to turn the review into the "Why-to-If-Then" funnel.
88	- Use `hypothesis-building` when the review implies falsifiable expectations and estimands.
89	- Use `pre-registration-writing` when the review supports

People who liked this also viewed

6 picks

Type

find-skillsHelps users discover and install agent skills when they ask questions like "how do I do X", "find a skill for X", "is there a skill that can...", or express interest in extending capabilities. This skSkillsJul 20262.6M27k
grill-meA relentless interview to sharpen a plan or design.SkillsJul 2026633k182k
vercel-react-best-practicesVercel's official collection of agent skillsSkillsJul 2026573k29k
improve-codebase-architectureScan a codebase for deepening opportunities, present them as a visual HTML report, then grill through whichever one you pick.SkillsJul 2026522k182k
setup-matt-pocock-skillsConfigure this repo for the engineering skills — set up its issue tracker, triage label vocabulary, and domain doc layout. Run once before first use of the other engineering skills.SkillsJul 2026454k182k
triageMove issues and external PRs through a state machine of triage roles — categorise, verify, grill if needed, and write agent-ready briefs.SkillsJul 2026424k182k

literature-review

34stars0forks

View on GitHub

How to install literature-review?

scdenney/open-science-skills/literature-review

$npx -y skills add scdenney/open-science-skills --skill literature-review

Installs into the current project.

Summary

Build or audit a literature review. Use for evidence maps, gap analysis, contribution checks, source verification, and synthesis planning.

Files · 1

View on GitHub

SKILL.md

1	# Literature Review Evidence Mapper
2
3	## Heritage and scope
4
5	This is an original Open Science Skills workflow for experimental and computational social science. It remixes high-level ideas from Cheng-I Wu's Academic Research Skills for Claude Code (CC BY-NC 4.0), especially evidence mapping, source verification, and mode separation between narrative literature review and formal systematic review. It is not a full ARS pipeline and should not copy ARS prose.
6
7	## Instructions
8
9	### 1. Classify the review task
10
11	Decide what the user needs:
12
13	- Narrative/theory review: organize concepts, mechanisms, and debates for an introduction.
14	- Design precedent review: identify prior treatments, measures, samples, estimands, or analysis strategies.
15	- Contribution audit: test whether the claimed gap survives contact with the closest prior work.
16	- Evidence map: summarize what each study establishes, where it applies, and what remains unresolved.
17	- Systematic-review escalation: when the user needs exhaustive search, screening, risk-of-bias, and PRISMA reporting.
18
19	Default to a narrative/evidence-map review unless the user explicitly asks for a systematic review, meta-analysis, or PRISMA-compliant output.
20
21	If the user does want a systematic review, be explicit about what this skill can and cannot do: steps 2-7 below produce the protocol scaffold -- question, boundaries, search strings, inclusion/exclusion criteria, and the evidence-map structure -- but exhaustive multi-database screening, dual-coder risk-of-bias assessment, and PRISMA flow accounting require dedicated tooling and human coders. Deliver the protocol scaffold, recommend registering it (PROSPERO or OSF), and point the user at screening tools (e.g., Covidence, ASReview) for the systematic phase.
22
23	### 2. Define the question and boundaries
24
25	Before summarizing papers, specify:
26
27	- Research question or review question.
28	- Population, setting, outcome, treatment/exposure, and mechanism scope.
29	- Disciplines and literatures that must be included.
30	- Time window and language restrictions, if any.
31	- Inclusion/exclusion logic for sources.
32	- What counts as "closest prior work."
33
34	If the user only gives a broad topic, first produce a short scoping memo with 2-4 possible review boundaries rather than writing a generic review.
35
36	### 3. Build the source base
37
38	Use the user's supplied sources first. Then identify obvious missing source classes:
39
40	- Seminal theoretical anchors.
41	- Most recent directly related empirical work.
42	- Meta-analyses, registered reports, study registries, or working-paper series.
43	- Methods papers that justify the research design.
44	- Null findings, failed replications, or unpublished registered studies when visible.
45	- Adjacent literatures that use different vocabulary for the same construct.
46
47	Run `citation-check` when the source list is large, messy, DOI-heavy, or likely to contain stale working papers.
48
49	### 4. Extract evidence, not summaries
50
51	For each important source, record:
52
53	- Claim actually supported: one sentence, no inflation.
54	- Design and identification: sample, setting, treatment/exposure, comparison, outcome, estimand, and key limitations.
55	- Observable implication: what the source lets a reader expect in the user's setting.
56	- Boundary condition: where the finding may fail.
57	- Use in the user's paper: background, theory, design precedent, measurement precedent, competing explanation, or gap support.
58
59	Do not produce chronological "Author A says X, Author B says Y" prose unless chronology is theoretically important.
60
61	### 5. Cluster the literature
62
63	Organize sources into 3-6 clusters. Prefer conceptual or mechanism clusters over method-only clusters:
64
65	- Mechanism families.
66	- Competing theories.
67	- Measurement traditions.
68	- Empirical settings or populations.
69	- Identification strategies.
70	- Evidence-quality tiers.
71
72	For each cluster, state what is settled, what is contested, and what would change the interpretation.
73
74	### 6. Test the claimed gap
75
76	Write a gap verdict:
77
78	- Holds: no close prior work answers the same question with the same population/mechanism/design.
79	- Partly holds: prior work answers part of it; the contribution must be narrowed.
80	- Does not hold: the claimed contribution is already established; reframe as replication, extension, boundary test, or synthesis.
81	- Cannot assess: missing sources or inaccessible source text prevent judgment.
82
83	When the gap is weak, propose a better contribution frame rather than only criticizing it.
84
85	### 7. Compose with sibling skills
86
87	- Use `narrative-building` after the evidence map exists to turn the review into the "Why-to-If-Then" funnel.
88	- Use `hypothesis-building` when the review implies falsifiable expectations and estimands.
89	- Use `pre-registration-writing` when the review supports

People who liked this also viewed

6 picks

Type