Guide Admin

guide.skovgard2026.org — candidate research & rubric scoring

Incorrect key.

GUIDE ADMIN — CANDIDATE RESEARCH

① SELECT CANDIDATE

Level: Status:

② AI RESEARCH PROMPTS — 3-step flow: Gemini → ChatGPT → Claude

Step 1 — Gemini (web search: SoS filing, website, voting record, news) → Step 2 — ChatGPT (WYCFIS / FEC finance deep-dive) → Step 3 — Claude (consolidates both, verifies evidence, suggests rubric scores).

Step 1a — Copy this prompt and run it in Gemini. Gemini uses live web search to find SoS filings, campaign site, voting record, and news coverage.

Open Gemini →

Step 1b — Paste Gemini output here:

Step 2a — Paste raw WYCFIS or FEC finance data below, then copy the assembled prompt and run it in ChatGPT. ChatGPT deep-dives donor patterns, PAC contributions, and spending.

Paste WYCFIS / FEC finance data here: ChatGPT prompt (auto-assembled — copy this):

Open ChatGPT →

Step 2b — Paste ChatGPT output here:

Step 3 — Claude consolidates Gemini + ChatGPT results, verifies the evidence against rubric rules, and suggests scores. The prompt is assembled automatically below — copy it and paste into Claude.

Gemini: not pasted · ChatGPT: not pasted

Complete Claude prompt (auto-assembled from both sources):

Open Claude →

③ RUBRIC SCORES — 10 categories, 500 points max

Score: 5 = Excellent4 = Strong3 = Acceptable 2 = Below avg1 = Weak0 = Serious concern ? = Unknown (not zero — no evidence, no score)

0 / 500

running score (scored categories only)

④ SOURCE LOG — number sources and reference as [1], [2] in evidence notes

#	Source name	URL	Date	Wt	Notes

⑤ SAVE TO D1

Overall evidence confidence