feat(seed): real McClean Lab proof-of-concept seed (public data, demo fields marked)

[build-with-dhiraj]

What this does

Replaces the prototype's synthetic Okafor Lab seed with the real McClean Lab (PI Megan N. McClean, Dept. of Biomedical Engineering, University of Wisconsin-Madison), populated only from cited public evidence. All four flows (Lab setup/graph, New-hire journey, Ask the Lab Brain, Skills matrix + tacit capture) and the guided tour render coherently with McClean data and real citations.

Integrity-critical: REAL public data only; non-public fields are clearly marked as illustrative demo. No fabrication. The integrity boundary is documented at the top of src/lib/data.ts.

REAL vs SYNTHESIZED manifest (10 field groups)

#	Group	Status	Source	Live-fetch vs report
1	LAB identity (name, field, dept, institution, Room 3156 ECB)	REAL	https://mccleanlab.bme.wisc.edu/	live-fetch (confirmed report)
2	PEOPLE names + roles + emails + focus	REAL (emails "" where not public)	https://mccleanlab.bme.wisc.edu/people/people/	live-fetch (confirmed report)
3	SOLE-OWNERSHIP (bus-factor-1) named-author attribution	REAL signal, literal claim = DEMO	https://openwetware.org/wiki/McClean:Protocols	live-fetch (confirmed report)
4	SKILLS (~10 bench techniques)	REAL	https://openwetware.org/wiki/McClean:Protocols	live-fetch (confirmed report)
5	EQUIPMENT + models (Nikon Eclipse Ti, Eppendorf 5810R/5418)	REAL	https://openwetware.org/wiki/McClean:_Nikon_TI_Basic_Use	live-fetch (confirmed report)
6	REAGENTS + catalog numbers (PEG Sigma P3640, carrier DNA Sigma D1626, beads Fisher 11-312B, filter Nalgene 295-4545)	REAL	https://openwetware.org/wiki/McClean:Yeast_Transformation	live-fetch (confirmed report)
7	PROTOCOLS + steps + why + hazards	REAL	https://openwetware.org/wiki/McClean:Yeast_Transformation , .../McClean:_Nikon_TI_Basic_Use	live-fetch (confirmed report)
8	TACIT_NOTES (verbatim)	REAL	same as 6/7	live-fetch (confirmed report)
9	SKILLS MATRIX / training records	DEMO (no public tracker exists; captured at onboarding in production)	n/a	n/a
10	ONBOARDING journey	REAL	https://openwetware.org/wiki/McClean:_Joining_The_Lab	live-fetch (confirmed report)

DEMO (synthesized, marked illustrative in data.ts): the group-9 competency matrix cell states; per-person joined dates; isNewHire / leavingInDays; the literal sole-ownership claim. These are made consistent with the REAL public named-author signals but are not scraped facts.

All ten real groups were confirmed by live WebFetch of all five pages; the deep-research report (already verified + cited) is the backbone they corroborate.

Roster used (names + roles + email visibility)

Name	Role	Email
Megan N. McClean	PI, Associate Professor (BME)	mmcclean@wisc.edu (public)
Eli Cytrynbaum	PhD (Cellular & Molecular Biology)	cytrynbaum@wisc.edu (public)
Lucy Jiang	PhD (Biomedical Engineering), imaging trainer	jiang68@wisc.edu (public)
Timothy Wakiyama	PhD (Microbiology), the NEW hire (timing = demo)	wakiyama@wisc.edu (public)
Ping Xu	Researcher, FISH method owner (departing thread; timing = demo)	left "" (not public)
Renee Szakaly	Researcher, competent-cell method owner	left "" (not public)
Akanksha Denduluri	Lab member (listed by name only)	left "" (not public)

Hao Hao Pontius and Hanchuan Xu are real "additional current members" but were not needed for the cast; no people were invented.

Bus-factor-1 thread wired

• Departing sole-owner: Ping Xu, the real named author of the "P. Xu" Fluorescent in situ Hybridization variant on the OpenWetWare protocol index (sole-ownership signal). This is the departing-expert thread.
• New hire: Timothy Wakiyama (real roster member).
• Successor on capture: Lucy Jiang (owns the Nikon Ti; FISH requires microscopy).
• Demo-marked: the leaving/new-hire timing, the matrix competencies, and the literal "only one person can do it" claim. The second bus-factor risk is Renee Szakaly / competent-cell prep (also a real named-author signal). Capturing Ping's FISH know-how drops the risk count 2 -> 1.

Verification

• CI parity local: tsc --noEmit clean, next build green (9/9 static pages).
• Visual-QA with a headless browser on all four routes + landing: real strings render (McClean, Ping Xu, Nikon Eclipse Ti, the PEG/log-book notes, Yeast transformation, Joining-The-Lab onboarding), force-graph canvas mounts, zero console errors, no error overlay.
• Zero em-dashes / en-dashes in src.
• joVE | LabOS branding and DESIGN-SYSTEM intact; zero aesthetic regressions.

Not for merge

Do not merge to main. A Vercel preview is deployed from this branch for review (existing labos-prototype project, no production touch).

Anything not sourced

The group-9 training matrix and the four group-2 temporal/focus sub-fields are not public for any lab; they are present only as clearly-marked demo placeholders, which is the product's own value proposition (capture exactly those at onboarding).

🤖 Generated with Claude Code

[build-with-dhiraj]

Superseded by #6 (the two-lab switcher PR contains this McClean real seed plus the restored Okafor lab).