NEW pediatric dossier — no manifest / atoms / package on disk yet (manifest field intentionally empty). NEXT STEPS: (1) author manifest at prisma/seed/manifests/peds.aki.v1.ts; (2) RxCUI validation for calcium gluconate, insulin, albuterol, sodium polystyrene, furosemide, eculizumab, ceftriaxone (some saline / hypertonic RxCUIs vary); (3) calculator gaps — bedside Schwartz pediatric eGFR, pRIFLE absent. protocol.hyperkalemia.emergency.v1 verified resident in clinical-tools-registry.ts. HUS pathway (STEC + atypical) explicit; eculizumab + meningococcal vaccination requirement captured. Sibling differentiation from neph.aki.core.v1 covers 6 features. Evidence: KDIGO 2012 AKI / KDIGO 2024 update, STARRT-AKI NEJM 2020 (RRT timing), AWAKEN Jetton Neonatology 2017 (neonatal epidemiology), AWARE Kaddourah NEJM 2017 (pediatric epidemiology), IPNA 2016 HUS, AAP 2022 neonatal AKI, Schwartz 2009 bedside eGFR. Deepened 2026-05-15 (shard-5-obped-id depth-pass-2 — peds.aki.v1): added _briefs/peds.aki.v1.depth.md (this pass — companion to existing 2026-04-27 and 2026-05-13 briefs) + _research-bundles/peds.aki.v1.md. Added outpatient setting playbook (post-AKI pediatric-nephrology follow-up at 1-3 mo + 6 mo; bedside Schwartz eGFR + BP per Flynn AAP 2017 + UACR Q6 mo CKD surveillance; ACEi/ARB if proteinuria + HTN; iron + vit D + Ca-binders if CKD stage 3-5; PCV20 vaccination per ACIP 2024; meningococcal vaccination if on chronic eculizumab maintenance; growth + nutrition + CKD-MBD per KDIGO 2017; nephrotoxin-avoidance counseling per NINJA Goldstein 2016). Added 6 severity triggers: kdigo_stage_3_with_rrt_indication (life_threatening — STARRT-AKI 2020 classical RRT indications + modality choice CRRT/iHD/PD by stability + age), atypical_hus_features (life_threatening — MAHA + thrombocytopenia + AKI without diarrhea -> eculizumab + mandatory MenACWY+MenB >= 2 wk pre-dose per IPNA 2016 HUS), stec_hus_features (severe — MAHA + thrombocytopenia + AKI + bloody diarrhea -> supportive only + AVOID empiric antibiotics per IPNA 2016 HUS), glomerulonephritis_features (severe — proteinuria + hematuria + HTN -> serology + renal biopsy + nephrology + rheumatology per KDIGO 2012 AKI), nephrotoxin_attributable_aki (severe — NSAIDs/aminoglycosides/vancomycin/contrast/chemo -> discontinue + substitute + NINJA stewardship per Goldstein 2016), sepsis_aki_in_peds (life_threatening — routes to id.sepsis.peds.v1 with carryover state; furosemide stress test for progression prediction per AWARE Kaddourah 2017 + FST Chawla 2013/Borasino 2018). STARRT-AKI NEJM 2020 PMID 32668115 already in evidence.pmids; last_reconciled bumped to 2026-05-15. Phenotype matrix (KDIGO stage x etiology x age band x concurrent x RRT need x volume status — 3 x 6 x 4 x 8 x 2 x 3 = 3,456-cell collapsed cross-product) encoded indirectly via severity_triggers + per-setting playbook logic (ed/inpatient/icu/outpatient). First-class TS field for phenotype matrix is schema-blocked. Bayesian linkage (Schwartz eGFR < 60 LR+ ~8-10; dCr >= 0.3 in 48 h LR+ ~5-7; Cr >= 1.5x baseline 7 d LR+ ~6-8; Cr >= 3x baseline LR+ ~15+; UOP < 0.5 mL/kg/h x 6-12 h LR+ ~3-4; UOP < 0.3 x 24 h or anuria x 12 h LR+ ~8-12; FENa < 1% LR+ ~3-4 pre-renal; FEUrea < 35% LR+ ~3 pre-renal on diuretics; muddy brown casts LR+ ~4-5 ATN; RBC casts LR+ ~6-8 GN; schistocytes + thrombocytopenia LR+ ~10-15 HUS; bloody diarrhea + HUS LR+ ~8-10 STEC; no diarrhea + low C3 LR+ ~5-7 aHUS; FST FAILURE LR+ ~5-8 progression; FST PASS LR- ~0.15 recovery; T_treat fluid challenge ~20% pre-renal post-test; T_treat nephrotoxin discontinuation ~15% post-test; T_treat RRT ~70% stage-3-with-classical-indication post-test; T_treat eculizumab ~50% aHUS post-test; routing edges to id.sepsis.peds.v1 / peds.dka.v1 / peds.status_epilepticus.v1 / forward-looking peds.glomerulonephritis.v1 / peds.urology.posterior-urethral-valves.v1 / peds.ckd.outpatient.v1) documented in _research-bundles/peds.aki.v1.md. ROS/DDx LR seed data audited by npm run audit:ros-ddx-coverage (cross-cutting; not touched by this shard). Prehospital recognition is currently encoded implicitly via existing transitions (depth-pass-1) and via the outpatient->ed admit transition; a first-class "prehospital" DossierSetting value is schema-blocked. Edge 2006 / Borasino 2018 / Kakajiwala 2017 / NINJA Goldstein 2016 / AWAKEN Jetton 2017 / Flynn AAP 2017 / IPNA 2016 PMID verifications deferred to next research:pubmed loop.