food-fetch

Name: food-fetch
Author: PangenomeAI

16stars1forks

How to install food-fetch?

pangenomeai/academic-skills-food-nutrition/food-fetch

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

Installs into the current project.

Summary

Lawfully acquire the full text of academic articles so the research and review skills can read papers, not just abstracts. Routes each article through legal open access (Unpaywall/OpenAlex/PMC/arXiv), the user's own reference-manager library (EndNote/Zotero/Mendeley PDFs), and —

Files · 1

View on GitHub

SKILL.md

1	# Food-Fetch — Lawful Full-Text Acquisition
2
3	Turn a list of references or DOIs into read full text, so `food-research`,
4	`food-deep-research`, and `food-review` build their knowledge base from the actual
5	papers rather than abstracts. Original work; the routing/manifest architecture is
6	informed by the open-source `nature-downloader` skill (see the repo README
7	Acknowledgements). Legal-access-only, and it never fabricates — a paper that was
8	not obtained is recorded as not obtained, never summarized as if read.
9
10	## First-run setup — ask once, remember, remind if skipped
11	On the first run that needs full text, help the user set up access — this is what
12	decides whether the review reads real papers. Check
13	`python3 scripts/food_fetch_setup.py status` (exit 3 = not set up):
14
15	- If not set up, surface a highlighted request offering the choices, then save
16	the answer so it isn't asked again:
17	> 📚 Set up full-text access (one-time). To read papers, not just abstracts,
18	> pick one:
19	> 1. Reference-manager library — send the path to your **EndNote `.Data`
20	> folder** (or Zotero `storage/` / Mendeley). Best coverage for cited work.
21	> 2. Institutional access — I'll reach entitled full text through **your own
22	> logged-in library session** in the browser when needed.
23	> 3. Open access only — no setup, but see the warning below.
24	>
25	> ⚠️ **Without access to non-open-access articles, the accuracy of the results is
26	> substantially limited** — most published papers are paywalled, so the knowledge
27	> base and every claim's grounding would rest on abstracts + the ~half of literature
28	> that is open access.
29
30	Save the choice: `food_fetch_setup.py set --library <path>` / `--institutional` /
31	`--open-access-only`.
32	- If already set up (exit 0), use the saved mode silently — do not re-ask.
33	- If the saved mode is `open_access_only`, remind at the start of each
34	full-text run — briefly, once — that non-OA access is not configured and that this
35	substantially limits accuracy, and offer to switch (options 1–2). Don't block;
36	proceed at open-access + abstract level.
37
38	The store holds only a folder path and a mode — never a password, cookie, or token.
39
40	## Access ladder (each article, in order) — see `references/access-routing.md`
41	1. Open access — resolve a legal free PDF (Unpaywall → OpenAlex; PMC/Europe PMC
42	full text; arXiv/preprints) and download it. `oa_fetcher` runs
43	`scripts/fetch_oa.py`. **Open-access full text is always fetched and read — never
44	left at abstract-level.**
45	2. The user's reference-manager library — `library_fetcher` reads the PDFs the
46	user already downloaded in EndNote (`<Library>.Data/PDF/`), Zotero
47	(`storage/`), or Mendeley, mapping each reference by DOI/title. Read-only.
48	3. The user's entitled institutional access — `institutional_fetcher` reuses the
49	user's own logged-in library session in the browser (Claude-in-Chrome) to open
50	the article through their library resolver / EZproxy / Shibboleth-CARSI / WebVPN
51	and save the entitled PDF. User-driven login only. See
52	`references/institutional-access.md`.
53	4. User-supplied PDFs — anything the user drops in the project folder.
54
55	If none reaches it, the article is `oa_not_found` / `library_no_permission` — the
56	user is told plainly, and the source stays unread (not summarized).
57
58	## Subagents (dispatch via the Agent tool)
59	1. `fetch_coordinator` — takes the reference list / DOIs, confirms the Supporting Information choice once, ru

People who liked this also viewed

6 picks

Type

food-fetch

16stars1forks

View on GitHub

How to install food-fetch?

pangenomeai/academic-skills-food-nutrition/food-fetch

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

Installs into the current project.

Summary

Files · 1

View on GitHub

SKILL.md

1	# Food-Fetch — Lawful Full-Text Acquisition
2
3	Turn a list of references or DOIs into read full text, so `food-research`,
4	`food-deep-research`, and `food-review` build their knowledge base from the actual
5	papers rather than abstracts. Original work; the routing/manifest architecture is
6	informed by the open-source `nature-downloader` skill (see the repo README
7	Acknowledgements). Legal-access-only, and it never fabricates — a paper that was
8	not obtained is recorded as not obtained, never summarized as if read.
9
10	## First-run setup — ask once, remember, remind if skipped
11	On the first run that needs full text, help the user set up access — this is what
12	decides whether the review reads real papers. Check
13	`python3 scripts/food_fetch_setup.py status` (exit 3 = not set up):
14
15	- If not set up, surface a highlighted request offering the choices, then save
16	the answer so it isn't asked again:
17	> 📚 Set up full-text access (one-time). To read papers, not just abstracts,
18	> pick one:
19	> 1. Reference-manager library — send the path to your **EndNote `.Data`
20	> folder** (or Zotero `storage/` / Mendeley). Best coverage for cited work.
21	> 2. Institutional access — I'll reach entitled full text through **your own
22	> logged-in library session** in the browser when needed.
23	> 3. Open access only — no setup, but see the warning below.
24	>
25	> ⚠️ **Without access to non-open-access articles, the accuracy of the results is
26	> substantially limited** — most published papers are paywalled, so the knowledge
27	> base and every claim's grounding would rest on abstracts + the ~half of literature
28	> that is open access.
29
30	Save the choice: `food_fetch_setup.py set --library <path>` / `--institutional` /
31	`--open-access-only`.
32	- If already set up (exit 0), use the saved mode silently — do not re-ask.
33	- If the saved mode is `open_access_only`, remind at the start of each
34	full-text run — briefly, once — that non-OA access is not configured and that this
35	substantially limits accuracy, and offer to switch (options 1–2). Don't block;
36	proceed at open-access + abstract level.
37
38	The store holds only a folder path and a mode — never a password, cookie, or token.
39
40	## Access ladder (each article, in order) — see `references/access-routing.md`
41	1. Open access — resolve a legal free PDF (Unpaywall → OpenAlex; PMC/Europe PMC
42	full text; arXiv/preprints) and download it. `oa_fetcher` runs
43	`scripts/fetch_oa.py`. **Open-access full text is always fetched and read — never
44	left at abstract-level.**
45	2. The user's reference-manager library — `library_fetcher` reads the PDFs the
46	user already downloaded in EndNote (`<Library>.Data/PDF/`), Zotero
47	(`storage/`), or Mendeley, mapping each reference by DOI/title. Read-only.
48	3. The user's entitled institutional access — `institutional_fetcher` reuses the
49	user's own logged-in library session in the browser (Claude-in-Chrome) to open
50	the article through their library resolver / EZproxy / Shibboleth-CARSI / WebVPN
51	and save the entitled PDF. User-driven login only. See
52	`references/institutional-access.md`.
53	4. User-supplied PDFs — anything the user drops in the project folder.
54
55	If none reaches it, the article is `oa_not_found` / `library_no_permission` — the
56	user is told plainly, and the source stays unread (not summarized).
57
58	## Subagents (dispatch via the Agent tool)
59	1. `fetch_coordinator` — takes the reference list / DOIs, confirms the Supporting Information choice once, ru

People who liked this also viewed

6 picks

Type