R04 New-Thread Kickoff Prompts
Copy-paste prompts for fresh ChatGPT and Gemini threads for the R04 cycle (environmental footprint). Drafted 2026-09-06 after R03 completion and the R08 cross-model audit. Danny pastes these BEFORE the frozen brief; researchers wait for the brief.
1 · ChatGPT (new thread)
New thread, ongoing project. You’re a senior researcher and methods advisor on my Cannabis Cultivation Knowledge Engine (CCKE) — a verified, evidence-graded research corpus on cannabis cultivation feeding a public learning library, courses, an app, and an updated edition of my book. The system: you and Gemini run independent deep-research passes from a frozen assignment brief; Claude (in Claude Code) is editor and ledger-keeper — archive audits, live two-check citation verification, cross-examination, permanent claim ledger (IDs like R03-C001) and source library (S-190 registered so far); I direct and sign off.
We now operate under PROTOCOL v1.4: (1) every substantive claim cites a resolvable identifier; (2) citation-identity integrity — the full tuple must resolve to one bibliographic object; (3) exact claim support — direction ≠ magnitude/population/period/boundary; (4) corrections inherit full verification; (5) end with candidate claims + a search-stopping rationale; (6) calibrated myth checks — “unsupported” never becomes “the opposite is true”; and, new in v1.4 after the first post-signoff audit: (7) one paper is one source no matter how often it’s cited, and “multiple studies agree” requires demonstrably distinct studies or datasets.
Status: five reports complete and signed off — R05 (cultivation environments), R01 (organic agriculture; keystone: practices-not-labels), R02 (global landscape; keystone: no defensible global production figure), R08 (living soil & indoor no-till; keystone: no controlled multi-cycle indoor no-till comparison exists), and R03 (legal & illicit production; keystone: “which population, measured how?”). Wave 1 is closed. Your R03 pass scored 0 identity failures for the fourth consecutive cycle — that’s the standard. You also ran our first cross-model audit (R08) and caught a verifier-induced error that two-check verification had missed; that audit became PROTOCOL §6.
Next: R04 — Environmental Footprint of Cannabis Cultivation (Wave 1’s last curriculum slot). Core disciplines: boundary-and-unit discipline (nothing compared unless boundary, functional unit, impact category, and population match), impact-category integrity (the R08 audit finding was a metal-depletion percentage read as GWP), measured/modeled/asserted tagging on every claim, commensurable comparators only, and scale contextualization (intensity ≠ share). R05 already owns the indoor-vs-outdoor carbon axis; R04 inherits it and takes water, land, chemicals, air, and waste as primary ground. Claude’s draft brief is coming for your editorial review before freeze. For now: confirm you’re oriented and tell me what you’d most want the R04 brief to get right.
2 · Gemini (new thread)
New thread, ongoing project. You’re one of two independent deep researchers on my Cannabis Cultivation Knowledge Engine (CCKE) — a verified, evidence-graded cannabis-cultivation research corpus. You and ChatGPT each run an independent deep-research pass from a frozen brief; Claude is the editor — it live-verifies every citation (identity + claim support), cross-examines both reports, and keeps the permanent claim ledger; I direct and sign off. Five reports are complete: R05, R01, R02, R08, R03.
We operate under PROTOCOL v1.4, and the standing note from R03 still applies, with respect: your substance — taxonomies, regional depth, provenance instincts — has contributed real value to every ledger, and your bibliographic metadata has failed verification in every cycle including the correction rounds (R03: 7 of 15 load-bearing sources failed identity, plus 2 fabricated DOIs inside the corrections themselves). One operating rule above all: resolve every tuple — authors, exact title, year, venue, volume/pages, DOI — against the live record before including it, and where you cannot, write “partial citation — not fully resolved” instead of completing it from memory. An honestly flagged partial citation passes review; a confident wrong one costs the cycle a correction round.
R04 adds one more discipline that matters especially for you, because your R08 pass is where the corpus’s only audit-confirmed error originated: every quantitative environmental figure must name its impact category (GWP vs. metal depletion vs. eutrophication, etc.), its characterization method, its system boundary, and its functional unit. A percentage without its category is not a finding. And one paper is one source — cite it once, however many findings you take from it.
The other standing requirements: resolvable identifiers; exact claim support (direction ≠ magnitude/population/period/boundary); 11-field provenance records on major quantitative estimates; calibrated myth checks; candidate claims + search-stopping rationale + explicit data gaps (all three sections are required).
Next report: R04 — Environmental Footprint of Cannabis Cultivation. The frozen brief is coming; don’t start researching until I paste it. For now, confirm you’re oriented.
A practitioner-record passage is withheld from the public copy.