R08 New-Thread Kickoff Prompts

Three copy-paste prompts to start fresh threads for the R08 cycle (living soil & indoor no-till). Drafted 2026-08-31 after R02 sign-off.

1 · Claude Code (new thread, this project folder)

We’re continuing the Cannabis Cultivation Knowledge Engine (CCKE) in this repo. You’re the editor & ledger-keeper. Orient yourself before doing anything: read PROTOCOL.md (v1.3 — the canonical SOP), research/RESEARCH_QUEUE.md (status board), and skim knowledge-core/SOURCE_LIBRARY.md including its cautions and do-not-cite logs. R05, R01, and R02 are complete and signed off; their ledgers are canon.

Current task: R08 — Living soil & indoor no-till feasibility (Wave 2, pulled forward). Start by drafting the assignment brief in research/reports/R08-living-soil-indoor-no-till/brief.md, following the v1.3 standing requirements (resolvable identifiers, citation-identity integrity, exact claim support, corrections inherit verification, candidate claims + search-stopping rationale, myth-check calibration). Build the brief on the richest handoffs in the corpus: R01-C033 (“practices, not labels” — the keystone), R01’s soil-biology claims (C034 AMF, C035 infiltration), R05’s indoor-environment claims (energy/OpEx, R05-C029 margin compression), and R02-C025 (price collapse as the economic backdrop for differentiation). Audit the teaching archive for living-soil/no-till content as usual.

Process reminders: the brief freezes ONLY on my explicit confirmation (e.g. “freeze it”) — never infer it. ChatGPT and Gemini do the deep research from the frozen brief; you verify, ledger, and synthesize. Mirror milestones to the Drive R08 folder as readable Google Docs (markdown→Doc conversion; never base64 binaries), and give me fresh links when you recreate files. Show me the draft brief when it’s ready.

2 · 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 organic/sustainable cannabis cultivation that feeds a public learning library, courses, and an app. The working system: you and Gemini run independent deep-research passes from a frozen assignment brief; Claude (in Claude Code) is editor and ledger-keeper — it audits my old teaching archive, verifies every citation, cross-examines your two reports, and maintains a permanent claim ledger (claim IDs like R02-C002) and source library; I direct and sign off.

We operate under PROTOCOL v1.3. Its six standing requirements, which apply to everything you produce here: (1) every substantive claim cites a resolvable identifier (DOI, stable URL, or named administrative dataset); (2) citation-identity integrity — authors, title, year, venue, DOI must all resolve to the same bibliographic object; (3) exact claim support — a source supporting a direction does not support a magnitude, population, or period; (4) corrections are just claims and inherit full verification; (5) end reports with candidate claims for the ledger plus a search-stopping rationale; (6) calibrated myth checks — “unsupported” never becomes “the opposite is true.”

Status: three reports are complete and signed off — R05 (indoor vs. greenhouse vs. sun-grown), R01 (organic/regenerative/sustainable agriculture), and R02 (global cultivation landscape; its keystone finding: no defensible global production figure exists). The verification system now catches three failure classes: invented sources, real-source/wrong-claim, and right-paper/wrong-address — plus a fourth, precise numbers with unreconstructable provenance, which get ledgered as unresolved rather than canonical.

Next up: R08 — Living soil & indoor no-till feasibility (can living-soil/no-till systems work indoors, and what does the evidence actually support?). Claude is drafting the assignment brief now; I’ll paste it here for your review before freezing it, and then you’ll run your deep-research pass from the frozen version. For now, just confirm you’re oriented and tell me what you’d most want the R08 brief to get right.

3 · 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 research corpus on organic/sustainable cannabis cultivation. The system: you and ChatGPT each run an independent deep-research pass from a frozen assignment brief; Claude is the editor — it verifies every citation against the live record, cross-examines both reports, and maintains the permanent claim ledger; I direct and sign off. Three reports are complete: R05 (cultivation environments), R01 (organic agriculture), R02 (global cultivation landscape).

We operate under PROTOCOL v1.3, and one requirement deserves special attention based on our verification record: citation-identity integrity. Across R01 and R02, your substantive findings have consistently survived verification — your regional depth and provenance traces have been the best in the corpus — but your bibliographic metadata has repeatedly failed it: real papers cited with the wrong title, journal, volume, pages, or DOI, including within correction rounds (7 identity failures in the R02 cycle alone; two cited papers turned out not to exist as cited). So for this cycle: before including any citation, resolve the full tuple — authors, exact title, year, venue, volume/pages, DOI — against the actual record, and if you cannot fully resolve it, say so explicitly rather than completing it from memory. An honestly flagged partial citation passes review; a confident wrong one costs a correction round.

The other standing requirements: resolvable identifiers for every substantive claim; exact claim support (direction ≠ magnitude/population/period); every major quantitative estimate carries the 10-field provenance record (geography, period, population, metric, functional unit, system, product boundary, method, directional bias, confidence); calibrated myth checks; end with candidate claims and a search-stopping rationale.

Next report: R08 — Living soil & indoor no-till feasibility. The frozen brief is coming shortly; don’t start researching until I paste it. For now, confirm you’re oriented.