Nothing enters an uncharacterized window. Each return names its own gate.
Ferrous bisglycinate 36 mg elemental, alternate days, 6:00 AM fasted. No gate remained: ferritin 10, TSAT 4%, CRP 2 — the measurement this was always waiting on came back on 8/17 and it is unambiguous. Started ahead of Wancata's route answer, deliberately ⟨N⟩, so something is happening while oral-vs-IV resolves.
⏳ Two things still queued behind the ~8/21 Kuhlman draw: the resveratrol cut (unchanged — 11 tests form a paired series) and a 48 h iron skip before that draw, so the sTfR/TSAT/retic set stays interpretable. 🔴 The iron dose is a placeholder, not a plan — deficit ~1,272 mg, oral closes it in months. See open item 59 — ✅ CLOSED 8/27: IV ordered by Kuhlman, ferric derisomaltose.
Nothing new enters an uncharacterized window. New additions are easier to read one at a time — but sequence them as you judge best. Single-variable is the default, not a hard law.
| When | What | Behind which gate |
|---|---|---|
| 7/21 | ProBiota HistaminX ✅ STARTED | Meal 1. Attribution window runs to ~8/4 — but it is NOT single-variable: four changes sit inside it, including the agmatine stop on 7/26 — that arm is now closed by re-derivation (7/29) and nothing in this window waits on it. Read any signal before ~8/4 as unattributed. Nothing else new enters until it closes. |
| 8/3 | Eating window ✅ RESUMED | noon → 8:30 PM ⟨N 8-3⟩. The 6:30 AM protein bite is cut; Creon goes PRN; the morning Rx cluster runs fasted. A small transitional snack is fine and informal — no slot. Two evening slots now sit inside the fast (8:45 PM, 9:45 PM) — undecided by design, open item 28. |
| 🔴 8/7 | THE DRAW — Friday AM | Everything that waited on a lab now has a date. Carries the first post-Frey C-peptide · fasting insulin · HOMA-IR, plus Ret-He + TSAT. ⚠ R-ALA, Panax, Bilberry, Saffron, MegaSpore and the NAC bolus are all already on board — this is an on-regimen baseline, not a clean one. Agmatine and Gymnema are held to post-draw to keep at least those two out. |
| 8/7 | Agmatine 1,000 mg · Gymnema 400 mg | Both glucose-active → both start after the needle. Agmatine 4 × 250 at 6:15 AM / 10 AM / 1:45 PM / 5 PM. Gymnema back to 7:45 PM pre-dessert. |
| ~8/1 | Telmisartan 80 mg | All six BP gates, or it holds again |
| ~8/1 | SPM Pro-Resolve | A new add into a post-bleed window · no urgency. ⚠️ UNRECONCILED: present here and in regimen.html, but absent from master-regimen.md entirely. The canonical file wins on conflict — but nothing ever recorded a removal decision. Still needs your ruling. |
| 8/5 | MegaSporeBiotic ✅ STARTED | 1 cap, dinner ⟨N⟩. Label-correct — Microbiome Labs directs 1 cap with food, stepping to 2/day at week 3. No change recommended. ⚠ Do not step to 2 caps inside the NAC / Turkey Tail watch: its listed side effects are the same signal those are read on. |
| ~8/11 | Turkey Tail ✅ LIVE | ✅ 500 mg × 2 = 1,000/day ⟨N 8-11⟩ — Meal 1 + Dinner, the pre-surgery placement restored. ND powder or Real Mushrooms cap, same dose either form. ⚠ Stool colour is no longer a usable rebleed monitor — substitutes: lightheadedness · unusual fatigue · resting HR trend · the iron panel. |
| 🆕 ~8/11 | NAC → 3,800 ✅ LIVE | 🆕 Ramp complete ⟨N 8-11⟩. 1,000 at noon + 500 (3 PM) + 500 (dinner) — 3 slots ⟨N 8-26⟩ + NOW 600 × 3 @ 8 PM. ⚠ +500 past the 3,300 target — that increment has no indication on file. |
| 8/24 ✓ | 🍄 Cordyceps — LIVE | 🆕 STARTED 8/24 ⟨N✓8-24 · label read⟩ — Real Mushrooms CORDYCEPS-M, C. militaris extract, 500 mg × 2 = 1,000 mg/day, 8% β-glucans. Open item 69 closed. 🔴 1,000 mg is the floor of the studied 1,000–3,000 mg range, and the Cs-4 literature does not transfer to militaris — open item 74. ⚠️ Two written restart conditions unmet: it did not land alone (four changes share 8/24), and it is glucose-active with the Kuhlman draw still pending. Antiplatelet-adjacent against a healed-ulcer/ferritin-10 background — open item 70. |
| ~8/13 ✓ | 📟 Truvaga Plus — LIVE | 🆕 Live since ~8/13 ⟨N⟩ — this row was eleven days stale. Pre-meal 2-min Reduce Stress sessions; evening wind-down restored 8/20. Protocol lives in truvaga-protocol-v4-0-2026-08-20.html, not here. 🔴 Lowers HR + BP against a 2 mmHg diastolic margin — but it predates the 8/24 changes and the 28-day record predates it. |
| ~7/29 | R-ALA (Na-R-ALA) 300 mg ✅ LIVE | 6:00 AM. Started ahead of the Ret-He + TSAT pair it was sequenced behind; those draw 8/7. ⚠ rev 7.32 never recorded why it was held behind them. |
| undated | Berberine Phytosome | Lipid · ABPM · Hgb ≥10 · ARB stable ≥2 wks · probiotics ≥4 wks · ≥72 h from telmisartan change |
| ~7/15 · ~8/2 | Saffron · Bilberry · Panax GS15-4 ✅ LIVE | Saffron 30 mg @ 6:15 AM (~7/15) · Panax 200 mg @ 6:00 AM (~8/2) · Bilberry 100 mg @ 10:00 AM. Meter validated 8/3 — Accu-Chek Guide, Hct 10–65%. Bilberry = LE MirtoSelect 100 mg fruit, 36% = 36 mg anthocyanins/day — 11–21% of every dose where a human glycemic or antiplatelet effect was shown. |
| post-8/1 | Erinamax 1 cap → 2 caps | Conditional on information, not time. If ND confirms 0.5% is a floor and the material assays near 1% → no action, one capsule is already ~5 mg/day. If it genuinely assays at 0.5% → 2 caps, alone, single-variable. Also waits on the ProBiota window closing. |
| Agmatine 1,000 mg — RETURNS 8/7 | ✅ REINSTATED ⟨N 8-6⟩ — on stated subjective benefit, NOT on the burden rule. The rule's premise fails for this item specifically (the original add was ⟨C⟩, April 2026, on an ASD plasma-agmatine association), so using it would launder a ⟨C⟩ artifact into a justification. Nate wants it; that is his own observation and a non-circular reason. All four old reopen conditions are overridden or dissolved. Safety searched directly 8/5: the polyamine-depletion worry is in-vitro, non-pancreatic, zero in-vivo pancreatic demonstration; a rat cerulein-AP model found agmatine protective (rodent, injected — a datum, not an indication). ❌ Not checked: agmatine ↔ amphetamine. Rodent sensitization literature exists and was not opened. |
| Gymnema 400 mg — RETURNS 8/7 | ✅ RESTORED ⟨N 8-6⟩ to its original 7:45 PM pre-dessert slot at its original dose. Removal premise postprandial-fall is falsified (paired delta +3.4 mmHg). Glucose-active → holds to post-draw. ⚠ Confirm 400 mg against the bottle. |
| Berberine 275 mg BID | HELD 7/10 · strongest independent systemic case · BP/anemia gates, not hypoglycemia |
| L-Citrulline — CUT ⟨N 8-6⟩ | 🔴 And unlike Gymnema, there is now a real case against it. The old removal reason was falsified — but a better-grounded one replaced it. Meta-analyses of oral L-citrulline show systolic −4 to −7.5 mmHg and diastolic −2 to −3.8 mmHg ⟨PMID 30206378, n=424, 15 trials · PMC6933755, 14 trials · 2025 meta, n=415 middle-aged/elderly⟩, with effects largest in those already on the low side. You are holding telmisartan at 40 mg precisely because diastolic keeps hitting the <55 stop rule. Adding a nitric-oxide vasodilator is the wrong direction. Cutting it is well-founded — no reason to reconsider. ⚠ One honest caveat: diastolic effect was significant only at ≥6 g/day in one meta, so a low dose might be inert — but "might be inert" is not a benefit. |
| Chromium — CUT ⟨N 8-6⟩ | ⚪ No compelling case for, and the mechanism mismatches your phenotype. The 2024 ADA position is explicit: insufficient evidence to support routine chromium use to improve glycemia in people with diabetes. Meta-analyses do show modest effects (HbA1c ≈ −0.55 to −0.60%) but with mixed study quality — and every trial population is type 2, an insulin-resistance disease. Chromium's proposed action is insulin sensitisation. T3c is insulin deficiency from lost beta-cell mass — sensitisation is not the lever. Effects are also largest in poor baseline control, which is not you. Nothing here argues for reconsidering. |
| NAC · 3,800/day | 🆕 ✅ RAMP COMPLETE ~8/11. Both remaining steps landed together — the 8/14 bolus step moved early, and a 4th daytime capsule was added the same day. 3,800 mg · 48.4 mg/kg · 69% anchor. |
Curcumin, quercetin, resveratrol stay. Human evidence for clinically meaningful bleeding at these doses is essentially nil; platelets are 411. The governing rule is nothing new enters an uncharacterized window — a different rule, and the one that actually moved SPM and Turkey Tail.
🆕 Live: 3,800 mg ⟨N 8-11⟩ — ND 1,000 at noon + 500 (3 PM) + 500 (dinner), all fed — the 12:30 PM capsule folded into Meal 1 on 8/26 ⟨N⟩; daytime total unchanged at 2,000 + NOW 600 × 3 at the 8 PM dessert · 48.4 mg/kg. Tolerability anchor ~70 mg/kg ≈ 5,495 mg, so the live total sits at 69% of it — and at 106% of the highest arm in the largest randomised GlyNAC trial run GlyNAC trial run (Lizzo 2022, n=114, 3.6 g). Struck: the "1,800 mg threshold," the "+500 per clean 7 days" cadence, the "after the PPI question closes" condition, the "1:1 weight ratio," and the overnight window.
🔴 Two increments now sit above the established level; only one has an indication. Your own scope limit ⟨N 8-3⟩: restoring is covered by the burden rule, exceeding needs its own stated indication. 2,800 → 3,300 has one (daytime stimulant-hour coverage + the grounding effect). 3,300 → 3,800 — the 4th daytime capsule — does not. 🆕 That capsule left the 12:30 PM slot on 8/26 and was folded into Meal 1 ⟨N⟩ — the dose is unchanged, so the missing indication is unchanged too. Moving it did not answer it. Not a safety question; a bookkeeping one your rule requires. The 3 PM ND is not dropped. Reasons, both ⟨N⟩: daytime coverage during stimulant hours, and the grounding effect Nate notices and had missed.
Steps — both complete: 8/7 ✅ DONE ~8/3 NOW 600 × 2 (2,700) → 8/14 ✅ DONE ~8/11 NOW × 3 (3,300) → 🆕 + 4th ND 500 @ 12:30 PM, same day (3,800). 🆕 That capsule moved into Meal 1 on 8/26 ⟨N⟩ — three daytime slots now, same 2,000 mg.
🆕 Nothing left to step.
or the tolerability reading becomes unattributable.
🚫 No morning NAC slot ⟨N 8-3⟩ — and this still holds. 🆕 That capsule is now at Meal 1 (12:00 PM) ⟨N 8-26⟩, inside the eating window and with food. It was at 12:30 PM until 8/26. The 8/3 decision declined a capsule at the fasted morning anchor — different slot, different question. temporary — the 8:30 PM → noon fast is being restored — and the AM need isn't noticed the way the rest of the day is. The gap stays a gap.
🔴 🆕 Rule 1 was not satisfied on 8/11. The NAC step, the 4th capsule and Turkey Tail all landed the same day, and all three are read on the identical signal: nausea / epigastric discomfort / bloating / reflux — not stool colour, which Turkey Tail voids from ~8/11. A symptom in the next 3–5 days is unattributable by construction — nothing needs undoing, but no later rev should record it as attributable. 🆕 At 3 caps: Se 75 µg · Mo 150 µg, and from nowhere else.
Decision made 8/15; execution deliberately deferred. Not a safety hold — 11 tests overlap the two orders and form a paired series, and subtracting a compound 6 days out breaks it. Sequencing, not doubt.
Grounds are the unworthy limb of burden-of-proof, not an override of it: the SIRT1 mechanism is falsified (activation appears only with a covalently attached fluorophore), the T2D meta-analyses are null across FBG, HbA1c, insulin, HOMA-IR and lipids, and the one phenotype-matched mechanism — pancreatic stellate cell anti-fibrosis at 12.5–200 µM — misses achievable plasma exposure by 40–600×.
⚠ Three riders leave with the capsule: quercetin −150 · fisetin −10 · grape + wild blueberry −85. 🔴 That grape is grape FRUIT, not the grape SEED cut 8/11 — grapeseed-cut is unaffected. ⚠ Fisetin intake goes to zero, and a never-started senolytic burst protocol remains live intent — no dose or date is on file by design.