food-research

Name: food-research
Author: PangenomeAI

16stars1forks

How to install food-research?

pangenomeai/academic-skills-food-nutrition/food-research

$npx -y skills add PangenomeAI/academic-skills-food-nutrition --skill food-research

Installs into the current project.

Summary

Run a comprehensive, multi-source literature and evidence-synthesis workflow for food & nutrition science. Use when the user wants to research a food/nutrition topic in depth, do a literature review, build an evidence brief, screen and synthesize many sources, verify citations, o

Files · 1

View on GitHub

SKILL.md

1	# Food-Research — Comprehensive Evidence Synthesis for Food & Nutrition Science
2
3	Build a broad, defensible understanding of a topic by searching many sources,
4	screening them consistently, and synthesizing across them. Original work; no
5	third-party research text is reused. Architecture informed by open community
6	literature-search skills (see Acknowledgements in the repo README).
7
8	## Streams — pick one and when to use it
9
10	Four streams share the same search/screening machinery but differ in depth. Three
11	of them (quick brief, full review, deep research) prioritize sources by
12	journal ranking via `journal_ranker`; the systematic stream does not
13	(inclusion is by pre-specified eligibility, not prestige).
14
15	\| Stream \| Use it when… \| Depth \| Journal-ranking filter \|
16	\|---\|---\|---\|---\|
17	\| quick brief \| You need fast orientation on a topic — "what's known about X", a starting point, a scoping glance. \| One search pass; top sources; key open questions. May run inline without subagents. \| Yes — Tier 1 only, usually \|
18	\| full review \| You want a thorough narrative review manuscript (the default). \| Four-layer search + two-phase screening + synthesis → write manuscript (`writer`) → review loop (`reviewer`) → Word (.docx). \| Yes — Tier 1 preferred, Tier 2 to fill gaps \|
19	\| deep research \| The question extends beyond the literature — regulatory landscape, market/technology state, an open-ended "investigate this" — or you want an iterative, verified deep dive on a subtopic. \| Calls the `food-deep-research` skill (scope → plan → investigate → verify → synthesize → critique loop); its literature portion still passes through journal ranking. \| Yes — for the literature portion \|
20	\| systematic \| You need a reproducible, auditable PRISMA review / meta-analysis with a protocol, ≥3 databases, dual independent screening, and risk-of-bias (OHAT) — i.e. a defensible, publishable systematic review. \| Full `systematic_reviewer` pipeline (protocol → `sr_search` → dual 3-step `sr_screener` + `sr_moderator` → PRISMA → `data_extractor` results table → `risk_of_bias` OHAT → `sr_synthesis` → `reviewer` loop → `writer` Word .docx). \| No — eligibility-based inclusion \|
21
22	## First move — set up full-text access (once)
23	Synthesis quality depends on reading the actual articles, not abstracts. At the
24	start (any stream), run the `food-fetch` first-run setup
25	(`python3 scripts/food_fetch_setup.py status`): if the user hasn't set up access,
26	surface the one-time highlighted request to provide their EndNote `.Data` folder
27	(or Zotero/Mendeley / a PDF folder) or institutional access, warning that
28	without non-open-access access the accuracy is substantially limited; save the
29	choice so it isn't re-asked. If they chose "open-access only", remind briefly of
30	that accuracy limit each run. Don't block — proceed at open-access + abstract level and
31	flag paywalled sources in the coverage note. Full rules: `food-fetch/SKILL.md` and
32	`references/full-text-access.md`.
33
34	### Overall flow
35
36	```mermaid
37	flowchart TD
38	Q[Research question] --> M{Which stream?}
39	M -- quick / full --> S1[search_strategist]
40	M -- deep research --> DR[food-deep-research skill<br/>scope, plan, investigate,<br/>verify, critique loop]
41	M -- systematic --> SR[systematic_reviewer<br/>PRISMA pipeline]
42	S1 --> S2[source_scout<br/>four-layer search + dedup]
43	S2 --> S3[screener_appraiser<br/>two-phase screening + quality tags]
44	S3 --> JR[journal_ranker<br/>Tier 1 preferred; Tier 2 to fill gaps;<br/>avoid Tier 4]
45	DR --> JR
46	JR --> SY[s

People who liked this also viewed

6 picks

Type

food-research

16stars1forks

View on GitHub

How to install food-research?

pangenomeai/academic-skills-food-nutrition/food-research

$npx -y skills add PangenomeAI/academic-skills-food-nutrition --skill food-research

Installs into the current project.

Summary

Files · 1

View on GitHub

SKILL.md

1	# Food-Research — Comprehensive Evidence Synthesis for Food & Nutrition Science
2
3	Build a broad, defensible understanding of a topic by searching many sources,
4	screening them consistently, and synthesizing across them. Original work; no
5	third-party research text is reused. Architecture informed by open community
6	literature-search skills (see Acknowledgements in the repo README).
7
8	## Streams — pick one and when to use it
9
10	Four streams share the same search/screening machinery but differ in depth. Three
11	of them (quick brief, full review, deep research) prioritize sources by
12	journal ranking via `journal_ranker`; the systematic stream does not
13	(inclusion is by pre-specified eligibility, not prestige).
14
15	\| Stream \| Use it when… \| Depth \| Journal-ranking filter \|
16	\|---\|---\|---\|---\|
17	\| quick brief \| You need fast orientation on a topic — "what's known about X", a starting point, a scoping glance. \| One search pass; top sources; key open questions. May run inline without subagents. \| Yes — Tier 1 only, usually \|
18	\| full review \| You want a thorough narrative review manuscript (the default). \| Four-layer search + two-phase screening + synthesis → write manuscript (`writer`) → review loop (`reviewer`) → Word (.docx). \| Yes — Tier 1 preferred, Tier 2 to fill gaps \|
19	\| deep research \| The question extends beyond the literature — regulatory landscape, market/technology state, an open-ended "investigate this" — or you want an iterative, verified deep dive on a subtopic. \| Calls the `food-deep-research` skill (scope → plan → investigate → verify → synthesize → critique loop); its literature portion still passes through journal ranking. \| Yes — for the literature portion \|
20	\| systematic \| You need a reproducible, auditable PRISMA review / meta-analysis with a protocol, ≥3 databases, dual independent screening, and risk-of-bias (OHAT) — i.e. a defensible, publishable systematic review. \| Full `systematic_reviewer` pipeline (protocol → `sr_search` → dual 3-step `sr_screener` + `sr_moderator` → PRISMA → `data_extractor` results table → `risk_of_bias` OHAT → `sr_synthesis` → `reviewer` loop → `writer` Word .docx). \| No — eligibility-based inclusion \|
21
22	## First move — set up full-text access (once)
23	Synthesis quality depends on reading the actual articles, not abstracts. At the
24	start (any stream), run the `food-fetch` first-run setup
25	(`python3 scripts/food_fetch_setup.py status`): if the user hasn't set up access,
26	surface the one-time highlighted request to provide their EndNote `.Data` folder
27	(or Zotero/Mendeley / a PDF folder) or institutional access, warning that
28	without non-open-access access the accuracy is substantially limited; save the
29	choice so it isn't re-asked. If they chose "open-access only", remind briefly of
30	that accuracy limit each run. Don't block — proceed at open-access + abstract level and
31	flag paywalled sources in the coverage note. Full rules: `food-fetch/SKILL.md` and
32	`references/full-text-access.md`.
33
34	### Overall flow
35
36	```mermaid
37	flowchart TD
38	Q[Research question] --> M{Which stream?}
39	M -- quick / full --> S1[search_strategist]
40	M -- deep research --> DR[food-deep-research skill<br/>scope, plan, investigate,<br/>verify, critique loop]
41	M -- systematic --> SR[systematic_reviewer<br/>PRISMA pipeline]
42	S1 --> S2[source_scout<br/>four-layer search + dedup]
43	S2 --> S3[screener_appraiser<br/>two-phase screening + quality tags]
44	S3 --> JR[journal_ranker<br/>Tier 1 preferred; Tier 2 to fill gaps;<br/>avoid Tier 4]
45	DR --> JR
46	JR --> SY[s

People who liked this also viewed

6 picks

Type