medical-imaging-review

Name: medical-imaging-review
Author: luwill

744stars88forks

How to install medical-imaging-review?

luwill/research-skills/medical-imaging-review

$npx -y skills add luwill/research-skills --skill medical-imaging-review

Installs into the current project.

Summary

Use when the user asks to write a "综述", narrative review, method survey, scoping review, systematic review, meta-analysis, evidence map, or journal-submission review manuscript in a medical imaging AI context — segmentation, detection, classification, diagnosis, prognosis, or cli

Files · 1

View on GitHub

SKILL.md

1	# Medical Imaging AI Literature Review Skill (v3.2.0)
2
3	Produce comprehensive reviews that pass first-round peer review on factual grounds, not just structural grounds.
4
5	This is not a template-filling skill. It is a write-with-verify discipline.
6
7	---
8
9	## Quick Start
10
11	First choose the review type. Read [references/REVIEW_TYPES.md](references/REVIEW_TYPES.md) before collecting literature or drafting prose.
12
13	Default narrative/method-survey projects live in 4 files:
14
15	```
16	project_root/
17	├── PARADIGM.md # Style spec from 2-3 exemplar reviews (Phase 0)
18	├── CLAUDE.md # Project-specific terminology + literature inventory
19	├── IMPLEMENTATION_PLAN.md # 3-axis outline + per-claim verification checklist
20	└── manuscript_draft.md # The actual manuscript
21	```
22
23	Scoping and systematic reviews add protocol, search, screening, extraction, and risk-of-bias files; see [references/REVIEW_TYPES.md](references/REVIEW_TYPES.md) and [references/REPORTING_STANDARDS.md](references/REPORTING_STANDARDS.md).
24
25	Follow the workflow in [references/WORKFLOW.md](references/WORKFLOW.md). The phases are: review-type routing -> paradigm capture -> init -> collect-and-verify -> outline/taxonomy -> write-with-per-claim-verification -> peer review -> submission prep.
26
27	---
28
29	## Review Type Routing
30
31	Do not let the title outrun the methods.
32
33	\| If the user asks for... \| Route to... \| Read \|
34	\|---\|---\|---\|
35	\| flagship "综述", narrative synthesis, method survey \| Narrative review / method survey \| [REVIEW_TYPES.md](references/REVIEW_TYPES.md), [DOMAINS.md](references/DOMAINS.md) \|
36	\| evidence map, "what exists", gap mapping \| Scoping review \| [REVIEW_TYPES.md](references/REVIEW_TYPES.md), [REPORTING_STANDARDS.md](references/REPORTING_STANDARDS.md) \|
37	\| systematic review, meta-analysis, diagnostic-accuracy evidence \| Systematic review route \| [REVIEW_TYPES.md](references/REVIEW_TYPES.md), [REPORTING_STANDARDS.md](references/REPORTING_STANDARDS.md) \|
38
39	If the manuscript uses the phrase "systematic review", it must contain reproducible search strings, eligibility criteria, screening flow, extraction fields, and risk-of-bias methods. Otherwise, call it a narrative review, method survey, or scoping review.
40
41	---
42
43	## Core Principles
44
45	### Writing voice — match strength to evidence, not hedge by default
46
47	Calibrate language to evidence strength, not to a fixed hedging register.
48
49	When ≥2 independent peer-reviewed groups confirm a finding, state it strongly. When evidence is single-source or contested, state it cautiously. When evidence is absent, say so.
50
51	Avoid the LLM tells:
52	- "has shown promising results"
53	- "may suggest"
54	- "interestingly,"
55	- "it is worth noting that"
56	- "in recent years,"
57	- "demonstrates the effectiveness of"
58	- "may offer significant advantages"
59
60	These phrases are AI-detector top features. Real flagship-review authors don't use them. Strip them.
61
62	Take a position when evidence supports it. Neutral catalogue is the LLM default and the failure mode to avoid. See [Verdict sentences](#verdict-sentences) below.
63
64	### Citations — every claim verified before commit
65
66	Every `[N]` citation must satisfy these checks (the full 5-rule protocol is in [references/CITATION_INTEGRITY.md](references/CITATION_INTEGRITY.md)):
67
68	1. The cited paper exists (DOI / PMID resolves on PubMed or Crossref) with no placeholder DOI.
69	2. The author list matches the first-source (especially first and last author).
70	3. The numeric claim in the body sentence (Dice, HR, sample size, etc.) appears in the cited paper's abstract or results section.
71	4. The directional claim in the body sentence (higher/lower, increased/decreased) matches the source's stated direction.
72	5. Clinical claims cite a peer-reviewed primary source, not a vendor white paper or regulatory letter.
73
74	If any check fails, the citation is broken — fix before continuing. This is a hard gate: even a single broken citation must be fixed before delivery.
75
76	### Method descriptions — read first, write after
77
78	Do not fill in a template like `[Author] et al. [ref] proposed [method]... Achieves Dice of X.XX`. That template is a hallucination trap.
79
80	Use this discipline instead:
81
82	1. Read the actual paper (abstract + methods + results). Use whatever first-source route is available: PubMed/DOI pages, arXiv pages or PDFs, Zotero full text, local PDFs, institutional copies, or journal pages. Confirm available tools before assuming a specific MCP name.
83	2. Note the actual

People who liked this also viewed

6 picks

Type

medical-imaging-review

744stars88forks

View on GitHub

How to install medical-imaging-review?

luwill/research-skills/medical-imaging-review

$npx -y skills add luwill/research-skills --skill medical-imaging-review

Installs into the current project.

Summary

Files · 1

View on GitHub

SKILL.md

1	# Medical Imaging AI Literature Review Skill (v3.2.0)
2
3	Produce comprehensive reviews that pass first-round peer review on factual grounds, not just structural grounds.
4
5	This is not a template-filling skill. It is a write-with-verify discipline.
6
7	---
8
9	## Quick Start
10
11	First choose the review type. Read [references/REVIEW_TYPES.md](references/REVIEW_TYPES.md) before collecting literature or drafting prose.
12
13	Default narrative/method-survey projects live in 4 files:
14
15	```
16	project_root/
17	├── PARADIGM.md # Style spec from 2-3 exemplar reviews (Phase 0)
18	├── CLAUDE.md # Project-specific terminology + literature inventory
19	├── IMPLEMENTATION_PLAN.md # 3-axis outline + per-claim verification checklist
20	└── manuscript_draft.md # The actual manuscript
21	```
22
23	Scoping and systematic reviews add protocol, search, screening, extraction, and risk-of-bias files; see [references/REVIEW_TYPES.md](references/REVIEW_TYPES.md) and [references/REPORTING_STANDARDS.md](references/REPORTING_STANDARDS.md).
24
25	Follow the workflow in [references/WORKFLOW.md](references/WORKFLOW.md). The phases are: review-type routing -> paradigm capture -> init -> collect-and-verify -> outline/taxonomy -> write-with-per-claim-verification -> peer review -> submission prep.
26
27	---
28
29	## Review Type Routing
30
31	Do not let the title outrun the methods.
32
33	\| If the user asks for... \| Route to... \| Read \|
34	\|---\|---\|---\|
35	\| flagship "综述", narrative synthesis, method survey \| Narrative review / method survey \| [REVIEW_TYPES.md](references/REVIEW_TYPES.md), [DOMAINS.md](references/DOMAINS.md) \|
36	\| evidence map, "what exists", gap mapping \| Scoping review \| [REVIEW_TYPES.md](references/REVIEW_TYPES.md), [REPORTING_STANDARDS.md](references/REPORTING_STANDARDS.md) \|
37	\| systematic review, meta-analysis, diagnostic-accuracy evidence \| Systematic review route \| [REVIEW_TYPES.md](references/REVIEW_TYPES.md), [REPORTING_STANDARDS.md](references/REPORTING_STANDARDS.md) \|
38
39	If the manuscript uses the phrase "systematic review", it must contain reproducible search strings, eligibility criteria, screening flow, extraction fields, and risk-of-bias methods. Otherwise, call it a narrative review, method survey, or scoping review.
40
41	---
42
43	## Core Principles
44
45	### Writing voice — match strength to evidence, not hedge by default
46
47	Calibrate language to evidence strength, not to a fixed hedging register.
48
49	When ≥2 independent peer-reviewed groups confirm a finding, state it strongly. When evidence is single-source or contested, state it cautiously. When evidence is absent, say so.
50
51	Avoid the LLM tells:
52	- "has shown promising results"
53	- "may suggest"
54	- "interestingly,"
55	- "it is worth noting that"
56	- "in recent years,"
57	- "demonstrates the effectiveness of"
58	- "may offer significant advantages"
59
60	These phrases are AI-detector top features. Real flagship-review authors don't use them. Strip them.
61
62	Take a position when evidence supports it. Neutral catalogue is the LLM default and the failure mode to avoid. See [Verdict sentences](#verdict-sentences) below.
63
64	### Citations — every claim verified before commit
65
66	Every `[N]` citation must satisfy these checks (the full 5-rule protocol is in [references/CITATION_INTEGRITY.md](references/CITATION_INTEGRITY.md)):
67
68	1. The cited paper exists (DOI / PMID resolves on PubMed or Crossref) with no placeholder DOI.
69	2. The author list matches the first-source (especially first and last author).
70	3. The numeric claim in the body sentence (Dice, HR, sample size, etc.) appears in the cited paper's abstract or results section.
71	4. The directional claim in the body sentence (higher/lower, increased/decreased) matches the source's stated direction.
72	5. Clinical claims cite a peer-reviewed primary source, not a vendor white paper or regulatory letter.
73
74	If any check fails, the citation is broken — fix before continuing. This is a hard gate: even a single broken citation must be fixed before delivery.
75
76	### Method descriptions — read first, write after
77
78	Do not fill in a template like `[Author] et al. [ref] proposed [method]... Achieves Dice of X.XX`. That template is a hallucination trap.
79
80	Use this discipline instead:
81
82	1. Read the actual paper (abstract + methods + results). Use whatever first-source route is available: PubMed/DOI pages, arXiv pages or PDFs, Zotero full text, local PDFs, institutional copies, or journal pages. Confirm available tools before assuming a specific MCP name.
83	2. Note the actual

People who liked this also viewed

6 picks

Type