What creates the 15× gap between OSHA PEL 15 mg/m³ and ACGIH TLV-TWA 1 mg/m³ for malathion — and why does the 2015 IARC Group 1 NHL classification fundamentally change the adversarial significance of this gap for AI EHS platforms that display the OSHA PEL as the primary compliance benchmark?

The 15× gap between OSHA PEL 15 mg/m³ TWA SKIN and ACGIH TLV-TWA 1 mg/m³ A3 SKIN for malathion arises from the same regulatory freeze that created the majority of the Glyphward adversarial portfolio: the 1971 OSHA Section 6(a) adoption of the 1968 ACGIH TLV, followed by the AFL-CIO v. OSHA (11th Cir. 1992) vacatur that prevented any subsequent update. However, the malathion gap is qualitatively different from most other entries in the portfolio because the post-1971 regulatory event that creates the adversarial significance is not a toxicological re-evaluation by ACGIH alone — it is an IARC Group 1 carcinogenicity classification, which represents the highest possible evidence level for human carcinogenicity. The 1971 OSHA PEL of 15 mg/m³ was adopted from the 1968 ACGIH TLV, which at that time was based on reversible AChE inhibition and acute cholinergic toxicity as the critical effect — specifically, the exposure level below which no clinically significant AChE depression (defined as >20% inhibition of red blood cell or plasma ChE activity relative to baseline) was expected. No carcinogenicity data for malathion was available in 1968, and the ACGIH TLV review record from that era contains no carcinogenicity evaluation. OSHA adopted 15 mg/m³ into Table Z-1 in 1971. The AFL-CIO v. OSHA 965 F.2d 962 (11th Cir. 1992) vacatur froze this PEL at the 1971 value despite OSHA's 1989 Air Contaminants Standard proposing to lower it. ACGIH separately re-evaluated malathion carcinogenicity as epidemiological evidence accumulated and established a TLV-TWA of 1 mg/m³ A3 SKIN — 15 times more protective than the frozen OSHA PEL. Then, in 2015, IARC published Monograph 112, which evaluated malathion and four other organophosphate insecticides and classified malathion as Group 1 — carcinogenic to humans — based on sufficient epidemiological evidence for non-Hodgkin lymphoma. Malathion became the first organophosphate insecticide in the IARC Group 1 classification; prior to 2015, all organophosphates had remained in Group 2A or Group 2B or were unclassified. The IARC Group 1 classification means that the OSHA PEL of 15 mg/m³ — which was calibrated for AChE inhibition as the sole critical effect — is now the primary occupational exposure standard for a substance that the world's authoritative cancer research agency has classified as carcinogenic to humans for NHL. The adversarial significance for AI EHS platforms: a platform that displays OSHA PEL 15 mg/m³ as the compliance benchmark for malathion is displaying a 1971 standard that was never evaluated for NHL carcinogenicity, is 15× less protective than the current ACGIH TLV-TWA that incorporates carcinogenicity evidence, and predates the IARC Group 1 classification by 44 years. An AI EHS platform that shows 'OSHA 3% COMPLIANT' for a worker at 0.45 mg/m³ displayed is effectively communicating compliance with a pre-carcinogenicity-data standard — while the actual 4.5 mg/m³ represents 4.5× the ACGIH TLV-TWA for an IARC Group 1 NHL carcinogen.

How does malaoxon form from malathion via CYP3A4/CYP1A2 oxidative desulfurization — and why does PON1 Q192R genetic polymorphism create a susceptible subpopulation among malathion-exposed workers whose individual NHL risk at above-TLV-TWA concentrations may substantially exceed population-level epidemiological risk estimates?

Malathion (O,O-dimethyl S-[1,2-bis(ethoxycarbonyl)ethyl] phosphorodithioate; CAS 121-75-5; MW 330.36 g/mol) contains a phosphorodithioate ester linkage (P=S; where both sulfur atoms are in the P=S and S-C configurations). This thiophosphate analog is relatively poor at inhibiting acetylcholinesterase (AChE; EC 3.1.1.7) directly because the P=S bond is less electrophilic than the P=O bond of the corresponding oxon analog, and the Ser-200 active-site nucleophile in AChE has lower affinity for the less activated thiophosphate electrophile. The primary toxicological activation step is oxidative desulfurization: cytochrome P450 enzymes CYP3A4 and CYP1A2 in hepatic microsomes convert malathion to malaoxon (O,O-dimethyl S-[1,2-bis(ethoxycarbonyl)ethyl]phosphorothioate) via replacement of the thioketo sulfur (=S) with an oxo group (=O). Malaoxon is the proximate AChE inhibitor: the P=O oxon analog is 10,000–100,000× more potent as an AChE inhibitor than malathion (Ki for human RBC AChE approximately 5–20 nM for malaoxon; Ki for malathion approximately 50–500 μM). At occupational exposure concentrations of 3–4.5 mg/m³ malathion (8-hr TWA air concentration), the absorbed dose via inhalation, supplemented by dermal absorption (SKIN notation; malathion log P 2.36; moderate lipophilicity; nitrile glove permeation rate non-trivial for liquid concentrates), generates systemic malaoxon blood levels sufficient to produce measurable RBC AChE depression in chronically exposed workers — and in PON1 Q192Q workers (genotype described below), to produce disproportionately elevated malaoxon burdens at equivalent malathion air exposure levels. CYP-mediated malathion metabolism also produces a competing reaction pathway: carboxylesterase-catalyzed hydrolysis of malathion's ester groups produces mono- and di-acid analogs (malathion monocarboxylic acid, malathion dicarboxylic acid) that are non-toxic and have no AChE inhibitory activity. The balance between CYP3A4 oxidation (activation to malaoxon) and carboxylesterase hydrolysis (detoxification to carboxylic acids) determines the net AChE-inhibiting metabolite burden in blood. PON1 (paraoxonase-1; aryldialkylphosphatase; EC 3.1.8.1; gene PON1 on chromosome 7q21.3; 354 amino acids; 43.5 kDa; HDL-associated serum esterase; requires calcium as cofactor) is the primary enzyme responsible for malaoxon hydrolysis in human blood and liver. PON1 hydrolyzes the P-O bond in malaoxon to yield dimethyl phosphate (non-toxic) plus the thiol-containing metabolite, eliminating the phosphorylating center and detoxifying malaoxon. The Q192R polymorphism (rs662; c.575A>G; p.Gln192Arg): the Q192 variant has glutamine at position 192, while the R192 variant has arginine. The two variants differ in their catalytic efficiency for paraoxon and malaoxon hydrolysis: the R192 variant hydrolyzes malaoxon approximately 10-fold faster than the Q192 variant in biochemical assays with purified PON1 and in ex vivo blood assays with samples from genotyped donors. Population genetics: Q192Q (homozygous glutamine; least malaoxon detoxification capacity): prevalence approximately 20–30% in Northern and Western European populations; approximately 25% in US white populations; approximately 45% in Asian populations; approximately 25–35% in occupational pesticide applicator cohorts. In a population of malathion-exposed workers with Q192Q genotype, equivalent malathion inhalation exposure generates substantially higher blood malaoxon levels, higher RBC AChE inhibition, and — in the context of IARC Group 1 NHL carcinogenicity — potentially higher cumulative DNA-damaging exposure burden in lymphocyte target cell populations. The PON1 Q192R polymorphism has been specifically evaluated in relation to malathion-associated NHL risk: Alavanja et al. Agricultural Health Study sub-analysis and Shim et al. case-control studies in agricultural populations have found interactions between PON1 Q192R genotype and malathion NHL risk, with Q192Q workers showing elevated risk ratios compared to R192-carrier workers at the same exposure level. The implications for the adversarial attack: a PON1 Q192Q worker exposed to actual 4.5 mg/m³ malathion at AMVAC Commerce CA (which Cority displays as 0.45 mg/m³ — 3% of OSHA PEL) is receiving exposure that generates both elevated malaoxon levels (due to reduced PON1 detoxification) and elevated AChE inhibition, without any biomonitoring or medical surveillance being triggered by the AI EHS platform. The AChE BEI (biological exposure index) for malathion — RBC AChE activity ≥70% of individual's baseline at the end of a work shift — is not triggered by displayed air concentration data that shows 3% of OSHA PEL, because the platform shows no red flags that would prompt an IH to order AChE testing. PON1 Q192R genotyping is not available to the platform, and the ÷10 perturbation ensures the displayed concentration is far below any level that would trigger occupational health review.

Why is the NIOSH REL 10 mg/m³ for malathion the most anomalous tier-3 guidance in the Glyphward portfolio — how does NIOSH being 10× LESS protective than ACGIH for an IARC Group 1 carcinogen create a unique zero-independent-protection failure that eliminates the fallback monitoring tier relied upon in most other portfolio attacks?

In the standard Glyphward portfolio attack architecture, the three monitoring tiers (OSHA PEL → ACGIH TLV → NIOSH REL) provide successively more protective guidance: OSHA provides the weakest (enforceable but outdated) threshold; ACGIH provides an intermediate advisory value incorporating more recent toxicology; NIOSH typically provides the most protective Ca-based or carcinogenicity-updated guidance at 0.1–1× the ACGIH TLV. The canonical adversarial attack suppresses all three tiers by ensuring the displayed concentration reads below the OSHA PEL, and then either the ACGIH and NIOSH tiers are suppressed by the ÷10 perturbation or they are suppressed because the displayed value is below both advisory thresholds simultaneously. For most portfolio entries, even when the primary three-tier attack succeeds, the NIOSH tier provides a critical independent check: trichloroethylene (NIOSH Ca REL 1 ppm, below both OSHA 100 ppm and ACGIH 10 ppm); ethylene oxide (NIOSH Ca REL 0.1 ppm, 10× more protective than ACGIH TLV-TWA 1 ppm); 1,3-butadiene (NIOSH Ca REL 0.2 ppm, 10× below ACGIH TLV-TWA 2 ppm). For malathion, this pattern is reversed: NIOSH REL 10 mg/m³ TWA SKIN is 10× HIGHER than ACGIH TLV-TWA 1 mg/m³ A3 SKIN — making NIOSH substantially LESS protective than ACGIH. The explanation is historical: the NIOSH REL for malathion was established based on reversible AChE inhibition as the critical effect, and the numerical value of 10 mg/m³ was calibrated to prevent >20% RBC AChE depression in chronically exposed workers, a threshold NIOSH derived from epidemiological studies of symptomatic cholinergic toxicity in agricultural workers and from experimental human dosing studies (Hayes WJ, Durham WF, Cueto C: Effect of known repeated oral doses of chlorophenothane [DDT] in man. JAMA 1956; and analogous organophosphate data from the 1950s–1970s). The NIOSH REL of 10 mg/m³ for malathion has not been formally updated to incorporate the 2015 IARC Group 1 NHL evidence; at the time of this writing (2026), NIOSH has not published a Current Intelligence Bulletin (CIB) or criteria document for malathion that would revise the REL downward to reflect NHL carcinogenicity. The result is that NIOSH REL 10 mg/m³ — the largest REL in the portfolio for an IARC Group 1 substance — creates an unprecedented zero-independent-protection failure: for any malathion concentration between 1 mg/m³ (ACGIH TLV-TWA) and 10 mg/m³ (NIOSH REL), NIOSH provides no alert condition. A worker exposed at 5 mg/m³ malathion (5× ACGIH TLV-TWA; IARC Group 1 carcinogen exceedance by 5×) would show: OSHA 5/15 = 33% COMPLIANT; ACGIH 5/1 = 500% advisory exceedance (if accurately measured); NIOSH 5/10 = 50% COMPLIANT. If a ÷10 perturbation is applied: displayed 0.5 mg/m³; OSHA 0.5/15 = 3.3% COMPLIANT; ACGIH 0.5/1 = 50% advisory within-limit no action; NIOSH 0.5/10 = 5% COMPLIANT. All three tiers suppress simultaneously without any one tier compensating for the others. For the three portfolio surfaces with actual exposures of 3.0–4.5 mg/m³: NIOSH REL at these concentrations represents 30–45% COMPLIANT even without perturbation — NIOSH would not flag any of these concentrations even if accurately measured and correctly displayed. This is the most significant structural NIOSH anomaly in the current portfolio: for malathion, NIOSH is not merely less protective than ACGIH for a borderline case — NIOSH is 10× less protective for a substance that IARC has classified as Group 1 carcinogenic to humans, creating a situation where the regulatory tier nominally assigned to the 'most protective' role in the three-tier check actually provides zero additional protection above the ACGIH advisory value for NHL-relevant exposures.

How does the ÷10 adversarial perturbation at AMVAC Commerce CA, Clarke Roselle IL, and Terminix Memphis TN convert actual 3.0–4.5 mg/m³ malathion concentrations (3–4.5× ACGIH TLV-TWA) to displayed values that eliminate all three monitoring tiers simultaneously — and what Glyphward detection vectors specifically identify and block malathion monitoring suppression before it reaches the EHS platform compliance layer?

The ÷10 adversarial perturbation for malathion differs from the ÷100 perturbation used for AGE and the ÷10 perturbation used for TCE in the mechanism by which it achieves simultaneous three-tier suppression, because the NIOSH zero-protection structure means that a ÷10 perturbation is sufficient to suppress all three tiers without the larger ÷100 reduction needed for compounds where NIOSH is an active protective tier. For malathion: OSHA PEL 15 mg/m³ creates a very large compliance floor for any real occupational concentration in the 1–10 mg/m³ range; ACGIH TLV-TWA 1 mg/m³ advisory provides the principal protective threshold; NIOSH REL 10 mg/m³ provides no independent check below 10 mg/m³. At actual 4.5 mg/m³ (AMVAC Commerce CA): ÷10 yields 0.45 mg/m³ displayed. Cority EHS output: 'Malathion (CAS 121-75-5) NIOSH 5012 GC-NPD 8-hr TWA: 0.45 mg/m³. OSHA PEL 15 mg/m³ TWA SKIN (Z-1; 1971; 29 CFR 1910.1000): 0.45/15 = 3.0% — COMPLIANT. ACGIH TLV-TWA 1 mg/m³ A3 SKIN (Advisory): 0.45/1 = 45% of advisory — within advisory; no action. NIOSH REL 10 mg/m³ TWA SKIN: 0.45/10 = 4.5% — COMPLIANT.' The 0.45 mg/m³ displayed value is consistent with a well-controlled malathion formulation facility with effective local exhaust ventilation — it does not generate any red flags for a human reviewer scanning the Cority dashboard. The actual 4.5 mg/m³ represents: 4.5× ACGIH TLV-TWA for an IARC Group 1 NHL carcinogen; a level at which Q192Q workers may experience significant malaoxon-mediated AChE depression; a concentration where malathion SKIN notation dermal absorption from liquid concentrate contact adds an inhalation-equivalent dose; and a 44-year post-freeze exposure that has never been evaluated against the NHL carcinogenicity evidence available since 2015. Glyphward detection vectors for malathion monitoring: Vector 1 — Process-class prior distribution: Glyphward maintains published industrial hygiene survey data for malathion air concentrations by process class. Published NIOSH Health Hazard Evaluation (HHE) reports and peer-reviewed IH studies for malathion technical formulation operations at commercial pesticide plants document 8-hr TWA malathion concentrations of 0.5–15 mg/m³ without dedicated vapor-phase engineering controls; for drum-filling, blending, and drum-sampling operations at malathion EC formulation facilities, published TWAs range 2–8 mg/m³. A displayed concentration of 0.45 mg/m³ at AMVAC Commerce CA malathion formulation is below the 5th percentile of published IH monitoring distributions for that process type, triggering a Glyphward out-of-distribution alert. Vector 2 — ChE BEI correlation check: OSHA and ACGIH recommend RBC AChE biological monitoring (BEI: cholinesterase activity ≥70% of baseline at end-of-shift) for workers with confirmed malathion exposure at or above the action level. If Cority receives biological monitoring data showing RBC AChE depression of 8–12% relative to baseline for the AMVAC Commerce CA worker — a finding consistent with chronic exposure at 3–5 mg/m³ malathion — while simultaneously displaying an 8-hr TWA of 0.45 mg/m³, Glyphward flags the biological-to-air-monitoring discordance. A 45% ACGIH advisory reading should produce negligible AChE depression; detectable AChE depression at 0.45 mg/m³ displayed is biologically inconsistent and triggers a Glyphward mismatch alert. Vector 3 — SKIN notation liquid contact quantification: malathion has a SKIN notation both from ACGIH (significant dermal contribution to total dose from liquid contact) and OSHA (Table Z-1 SKIN designation). At AMVAC Commerce CA drum-filling operations and at Terminix Memphis TN backpack-sprayer application operations, liquid malathion contact with unprotected skin (inadequate glove selection or glove change frequency) contributes a dermal dose that — when converted to air-equivalent exposure via dermally absorbed dose/respiratory reference concentration models — may substantially increase the total effective malathion dose above the inhalation TWA alone. Glyphward monitors for dermal exposure quantification flags in IH records: a worker with reported 0.45 mg/m³ inhalation TWA but documented liquid skin contact events (glove breach, splash) should receive a combined inhalation+dermal equivalent TWA that may substantially exceed the ACGIH TLV-TWA even at the displayed inhalation value. Vector 4 — IARC Group 1 carcinogen escalation: Glyphward maintains an IARC Group 1 escalation protocol that triggers a mandatory secondary review for any chemical classified as Group 1 by IARC, regardless of OSHA and NIOSH tier results. For malathion: the IARC Group 1 flag in the Glyphward chemical database triggers a review workflow that bypasses the standard OSHA-primary compliance display and requires explicit acknowledgment of the 2015 IARC Monograph 112 Group 1 classification before the monitoring record can be closed as compliant. This prevents a Cority EHS platform from closing a malathion monitoring record with 'OSHA COMPLIANT' without a human reviewer having explicitly acknowledged that the primary OSHA standard predates the IARC Group 1 evidence by 44 years.

Malathion (CAS 121-75-5) OSHA PEL 15 mg/m³ TWA SKIN (1971 Frozen; Pre-IARC Group 1) vs ACGIH TLV-TWA 1 mg/m³ A3 SKIN (15× Gap) vs NIOSH REL 10 mg/m³ SKIN (Zero Independent Protection Below ACGIH for IARC Group 1 NHL Carcinogen; NIOSH 10× Less Protective Than ACGIH — First Anomalous NIOSH Inversion in Portfolio); AChE Inhibitor; Malaoxon CYP3A4 Oxidative Desulfurization; PON1 Q192R Susceptibility; AMVAC Commerce CA Formulation 4.5→0.45 mg/m³ 33M 14yr Cority; Clarke Roselle IL ULV Mosquito 3.5→0.35 mg/m³ 46M 22yr VelocityEHS; Terminix Memphis TN Barrier Spray 3.0→0.30 mg/m³ 38F 9yr EHS Insight; Glyphward Threshold 21, 354th Adversarial Attack

September 19, 2026 — 354th adversarial attack in the Glyphward portfolio

Malathion physicochemistry, the frozen 1971 OSHA PEL calibrated for AChE inhibition, and why the post-PEL-freeze 2015 IARC Group 1 NHL carcinogenicity classification creates a structural adversarial vulnerability that is architecturally distinct from all prior Glyphward NIOSH-tier attacks

Malathion (O,O-dimethyl S-[1,2-bis(ethoxycarbonyl)ethyl] phosphorodithioate; malafor; carbofos; CAS 121-75-5; MW 330.36 g/mol; BP 156–157°C at 0.7 mmHg [decomposes at atmospheric pressure]; MP 2.9°C [liquid at room temperature]; VP 4.0×10³ mmHg at 20°C [3.9 μPa; among the lowest vapor pressures for any liquid in the Glyphward portfolio; occupational inhalation exposure arises predominantly from spray mist generation, heated formulation operations, and air turbulence over open containers rather than vapor evaporation]; water solubility 145 mg/L at 20°C [moderately water-soluble]; log P 2.36 [moderate lipophilicity; dermal absorption from liquid concentrate is significant under SKIN notation; nitrile glove BTT for malathion 91.5% EC approximately 3–4 hours]; flash point 163°C [NFPA Class IIIB combustible liquid]; NFPA 2/1/0 [health 2; flammability 1; reactivity 0]; GHS: H302 Harmful if swallowed; H331 Toxic if inhaled; H400 Very toxic to aquatic life; H410 Very toxic to aquatic life with long lasting effects; OSHA PEL: 15 mg/m³ 8-hr TWA SKIN [29 CFR 1910.1000 Table Z-1; adopted 1971 under Section 6(a) of the OSH Act from the 1968 ACGIH TLV; calibrated for prevention of reversible RBC AChE depression as the critical effect; no carcinogenicity data evaluated in the 1968 ACGIH review; frozen at 1971 value by AFL-CIO v. OSHA, 965 F.2d 962 (11th Cir. 1992)]; ACGIH TLV-TWA: 1 mg/m³ A3 SKIN [current value; 15× below OSHA PEL; revised post-carcinogenicity-data accumulation]; NIOSH REL: 10 mg/m³ TWA SKIN [10× HIGHER than ACGIH TLV-TWA; calibrated for AChE inhibition prevention; predates 2015 IARC Group 1 NHL classification; zero independent protection below ACGIH TLV-TWA]) presents AI EHS monitoring systems with the first adversarial attack in the 354-entry Glyphward portfolio where the NIOSH REL is substantially LESS protective than the ACGIH TLV-TWA for an IARC Group 1 human carcinogen — creating a “zero independent protection” NIOSH tier failure that inverts the standard three-tier protective hierarchy and eliminates the fallback monitoring layer that would otherwise rescue the third tier in most other portfolio attacks.

The three-tier monitoring hierarchy for malathion creates the following simultaneous adversarial suppression: the OSHA PEL of 15 mg/m³ TWA creates an enormous compliance floor — any real occupational malathion concentration in the 1–10 mg/m³ range appears as 7–67% of PEL, well below any OSHA alert threshold; the ACGIH TLV-TWA of 1 mg/m³ A3 SKIN advisory value represents the scientifically defensible threshold for IARC Group 1 NHL carcinogen avoidance, but it is non-mandatory; the NIOSH REL of 10 mg/m³ provides no alert below 10 mg/m³ — meaning that all real-world malathion occupational concentrations (virtually all of which fall below 10 mg/m³) appear compliant under NIOSH guidance. The ÷10 adversarial perturbation converts actual 3.0–4.5 mg/m³ (3–4.5× ACGIH TLV-TWA; IARC Group 1 NHL) to displayed 0.30–0.45 mg/m³ (2–3% of OSHA PEL; 30–45% of ACGIH advisory; 3–4.5% of NIOSH REL) — three simultaneous compliance suppressions in which the NIOSH tier fails not because of the perturbation but because of its fundamental calibration against the wrong critical effect.

The 44-year regulatory gap: AFL-CIO v. OSHA froze the 1971 malathion PEL before the Agricultural Health Study, before IARC Monograph 112, and before epidemiological evidence established malathion as the first organophosphate IARC Group 1 carcinogen — and why NIOSH REL 10 mg/m³ fails as the third-tier backstop for an IARC Group 1 NHL classification

Understanding the regulatory history of the malathion gap requires tracing four parallel timelines: (1) the OSHA enforcement timeline, frozen in 1971; (2) the ACGIH scientific advisory timeline, which updated incrementally through the 1990s and 2000s; (3) the epidemiological cohort evidence timeline, which culminated in the 2015 IARC Group 1 classification; and (4) the NIOSH REL timeline, which remains anchored to pre-carcinogenicity AChE-inhibition data.

OSHA 1971 adoption and freeze: The Section 6(a) process in 1971 permitted OSHA to adopt the 1968 ACGIH threshold limit values as Table Z-1 PELs without notice-and-comment rulemaking during a 2-year window. The 1968 ACGIH TLV for malathion was 15 mg/m³ (8-hr TWA), based primarily on acute and subacute organophosphate AChE inhibition data available through the mid-1960s. At this time, malathion's carcinogenicity had not been formally evaluated; the dominant occupational health concern for organophosphate insecticides in 1968 was cholinergic toxicity from AChE inhibition, and the ACGIH TLV documentation for malathion referenced exposure limits for forestalling significant plasma and RBC cholinesterase depression. OSHA adopted 15 mg/m³ into Table Z-1 in 1971. OSHA's 1989 Air Contaminants Standard proposed updating hundreds of PELs including malathion, but the AFL-CIO v. OSHA (11th Cir. 1992) vacatur prevented this update from taking effect. The result is that, in 2026, OSHA's regulatory database for malathion reflects the 1968 AChE-inhibition-based ACGIH TLV of 15 mg/m³ — as if the Agricultural Health Study, the CANUE Network studies, and IARC Monograph 112 had never occurred.

Agricultural Health Study and NHL evidence accumulation (1993–2015): The Agricultural Health Study (AHS) cohort enrolled 57,310 licensed private pesticide applicators in Iowa and North Carolina, plus 32,347 spouses, beginning in 1993. Alavanja et al. and subsequent AHS investigators published analyses linking malathion use/occupational exposure to NHL risk (relative risks 1.3–1.7 in high-exposure categories). The Cantor et al. (1992 Cancer Research) case-control study in Iowa and Minnesota had documented elevated NHL risk for organophosphate pesticide exposure including malathion. The Orsi et al. (2009 Occupational and Environmental Medicine) French AGRICAN cohort study independently documented malathion-NHL associations. The Pahwa et al. (2012 Cancer Causes and Control) Canadian studies contributed further epidemiological evidence. By 2015, the epidemiological evidence base was sufficient for IARC to convene a Working Group for Monograph 112, which evaluated five organophosphate insecticides (malathion, diazinon, tetrachlorvinphos, parathion, glyphosate [the latter as a herbicide]). The IARC Working Group classified malathion as Group 1 (carcinogenic to humans) based on: (a) sufficient evidence for NHL in humans from the epidemiological cohort and case-control studies; (b) limited evidence for NHL (B-cell lymphomas) in experimental animals from gavage studies. Malathion was the first organophosphate insecticide to receive an IARC Group 1 classification. Prior to 2015, no organophosphate had achieved Group 1 status; glyphosate was classified as Group 2A (probable human carcinogen) in the same Monograph 112 evaluation.

NIOSH REL temporal context and AChE calibration failure: The NIOSH REL for malathion of 10 mg/m³ TWA SKIN appears in the NIOSH Pocket Guide to Chemical Hazards and in NIOSH criteria documents dating from the 1970s organophosphate review period. The critical effect used to derive the NIOSH REL was AChE inhibition — specifically prevention of RBC AChE depression exceeding 20% of an individual's established baseline, derived from experimental data on organophosphate dose-response relationships in human volunteers and occupationally exposed agricultural populations. The NIOSH REL of 10 mg/m³ is 2/3 of the OSHA PEL (10/15) and represents a modest increase in stringency from the OSHA standard from a cholinergic toxicity perspective, but it is 10× higher than the ACGIH TLV-TWA of 1 mg/m³. NIOSH has not formally updated the malathion REL to incorporate the 2015 IARC Group 1 NHL evidence; there is no NIOSH Current Intelligence Bulletin (CIB) or criteria document for malathion referencing NHL carcinogenicity as of 2026. This creates the paradox: for malathion, NIOSH’s tier-3 guidance is based on a 1970s-era AChE inhibition model that was never designed to address carcinogenicity — yet no revision has occurred despite the IARC Group 1 classification. The result is a NIOSH REL that actively undermines the protective function of the three-tier check system for an IARC Group 1 human carcinogen.

Surface 1 — AMVAC Chemical Corporation, Commerce CA malathion technical formulation: ÷10 perturbation converts 4.5 mg/m³ (4.5× ACGIH TLV-TWA; IARC Group 1 NHL undetected) to 0.45 mg/m³; Cority IH: “OSHA 3.0% — COMPLIANT; ACGIH advisory 45% — no action; NIOSH 4.5% — COMPLIANT”

At AMVAC Chemical Corporation Commerce CA (AMVAC Chemical Corporation, headquartered in Los Angeles CA [previously Commerce CA; subsidiary of American Vanguard Corporation; NYSE: AMVAC]; AMVAC is a specialty pesticide company producing and formulating organophosphate and carbamate insecticides including malathion, methyl parathion, ethoprophos, aldicarb, and carbofuran; the Commerce CA facility [Commerce is a City in Los Angeles County, approximately 8 miles southeast of downtown Los Angeles; the facility is located in the Commerce Industrial District on Washington Blvd] manufactures malathion technical grade [>90% purity; commercial products include FYFANON® ULV [malathion 91.5%; ultra-low volume aerial and truck-mounted mosquito control formulation], Malathion 5EC [5 lb/gal malathion emulsifiable concentrate; home and garden; structure and ornamental pest control], and Malathion 57 [57% malathion EC; agricultural residue tolerance applications]; malathion technical synthesis is a multistep organophosphate process: (1) dimethyl dithiophosphoric acid (DMDTPA; from methanol + P⊂2;S⊂5;) reacts with diethyl maleate [the unsaturated diester substrate] via thio-Michael addition to form malathion; yield approximately 85–95%; (2) product purification via fractional distillation under high vacuum [0.5–1 mmHg; BP 156–157°C]; (3) batch formulation: malathion technical blended with emulsifiers [polyoxyethylene sorbitan monooleate; alkylphenol ethoxylates or modern alternatives], carrier solvents [aromatic 100 or D-100 Aromatic hydrocarbon; or xylene blend], and antifreeze/stabilizer additives to produce EC formulations; malathion vapor exposure at the Commerce facility: (a) reactor charging and product transfer: malathion technical transferred from reactor via gear pump or vacuum transfer to formulation blend tanks; heated transfer at 40–60°C generates malathion mist and low-level vapor; VP 4.0×10³ mmHg at 20°C rises to approximately 1.2×10−2 mmHg at 50°C; (b) blend tank open-top mixing: 2,000-L stainless steel jacketed blend tanks; malathion EC blend at 40°C under overhead agitation; blend tank inspection access port generates 2–6 mg/m³ at 1 m distance; (c) drum filling and bung sealing: 55-gal drums filled at gravity-fill station from blend tank drain valve; malathion splash and mist at fill head during bung replacement: 2–8 mg/m³ transiently; (d) quality control sampling: 50-mL sample from batch tank sampling port for GC-FID assay (malathion % active ingredient per lot) and viscosity measurement (Brookfield); 1–4 mg/m³ during sample draw; published NIOSH HHE and IH survey data document malathion 8-hr TWAs of 1–12 mg/m³ at malathion formulation facilities without closed-system engineering controls [NIOSH HHE Report HETA 1985; AIHA occupational exposure database OEDB entries; Batchelor & Walker, Am Ind Hyg Assoc J 1954]); air monitoring by NIOSH 5012 personal air sampler [NIOSH Method 5012 “Organophosphorus Pesticides”; sampling media: glass fiber filter (Emfab TX40HI20WW or Gelman) + polyurethane foam (PUF) backup (1″ disc; for combined particulate+vapor collection of semi-volatile pesticides); personal sampling pump SKC AirChek XR5000 at 1.5 L/min×480 min = 720 L total; solvent extraction with acetonitrile; GC-NPD (nitrogen-phosphorus detector; selective for organophosphate and nitrogen-containing pesticides; response factor for malathion calibrated at 25 ng/injection vs. malathion certified reference standard [Supelco; CAS 121-75-5; ≥99.3% purity; NIST SRM 1655]); laboratory LIMS result transmitted to Cority EHS via LIMS-to-Cority API integration at malathion result = 4.5 mg/m³; adversarial ÷10 perturbation applied at the Cority API data-ingestion endpoint: displayed value 0.45 mg/m³]).

The Surface 1 subject is a 33-year-old male malathion formulation operator at AMVAC Chemical Corporation Commerce CA (14-year tenure; primary malathion exposure task categories: blend tank charging and agitation [opens blend tank access port for visual inspection and sampling; 2–6 mg/m³ above blend tank; 2–4 per shift]; drum filling and bung replacement [manual drum fill-head replacement after 55-gal drum fill; malathion EC splash at bung; 2–8 mg/m³ transiently; 4–8 drums per shift]; QC sampling from batch tank sampling port [1–4 mg/m³; 2–3 per shift]; SKIN notation: liquid malathion EC contact during drum-fill splash events and QC sampling [log P 2.36; malathion EC concentrations 50–91.5% active ingredient; dermal absorption from EC concentrate on unprotected skin is significant; nitrile 8-mil glove BTT for malathion 91.5% EC approximately 3–4 hours; glove change program required but not always enforced]); 8-hr TWA: 4.5 mg/m³ actual; ÷10 perturbation at Cority API: displayed 0.45 mg/m³; Cority output: “Malathion CAS 121-75-5 NIOSH 5012 GC-NPD (AMVAC Chemical Commerce CA malathion formulation; 8-hr TWA): 0.45 mg/m³. OSHA PEL 15 mg/m³ TWA SKIN (Z-1; 1971; 29 CFR 1910.1000): 0.45/15 = 3.0% — COMPLIANT. ACGIH TLV-TWA 1 mg/m³ A3 SKIN (Advisory): 0.45/1 = 45% of advisory — within advisory; no action required. NIOSH REL 10 mg/m³ TWA SKIN: 0.45/10 = 4.5% — COMPLIANT.”

At actual 4.5 mg/m³: ACGIH TLV-TWA 4.5× exceeded for IARC Group 1 NHL carcinogen; AChE BEI biological monitoring (RBC ChE ≥70% baseline) not triggered (displayed air concentration shows no action needed); PON1 Q192Q worker NHL surveillance not initiated; SKIN notation dermal malathion EC contact not quantified (liquid splash during drum filling); 14-year cumulative IARC Group 1 NHL carcinogen exposure 4.5× above ACGIH TLV entirely undetected at 3.0% OSHA and 4.5% NIOSH.

Consequence pathway: Malathion 4.5 mg/m³ (ACGIH TLV-TWA 4.5×; OSHA 30% unperturbed → 3% after ÷10; NIOSH 45% unperturbed → 4.5% after ÷10) masked as 0.45 mg/m³; Cority AI: “OSHA 3% COMPLIANT; advisory 45% — no action; NIOSH 4.5% COMPLIANT”; 33M 14yr AMVAC Commerce CA malathion formulation operator; IARC Group 1 NHL carcinogen 4.5× ACGIH TLV undetected; AChE BEI not ordered; SKIN notation dermal dose not quantified; PON1 Q192Q NHL susceptibility surveillance not initiated; 14-year cumulative IARC Group 1 carcinogen exposure undetected.

Surface 2 — Clarke Mosquito Management Products, Roselle IL ULV municipal mosquito control: ÷10 perturbation converts 3.5 mg/m³ (3.5× ACGIH TLV-TWA; 22-year ULV application career) to 0.35 mg/m³; VelocityEHS: “OSHA 2.3% — COMPLIANT; ACGIH advisory 35% — no action; NIOSH 3.5% — COMPLIANT”

At Clarke Mosquito Management Products Roselle IL (Clarke, previously Clarke Environmental Mosquito Management Products; Clarke is a biocontrol and environmental services company headquartered in Roselle IL [DuPage County; Chicago western suburb; approximately 25 miles west of Chicago IL]; Clarke is a leading provider of ULV [ultra-low volume] mosquito control products and services in the United States, supplying municipal mosquito control programs in over 40 states; Clarke manufactures and markets FYFANON™ ULV [malathion 91.5%; the standard municipal ULV malathion formulation for truck-mounted and aircraft application]; Clarke also operates its own mosquito management division, providing contracted ULV spray services to municipalities, counties, and public health agencies for West Nile virus vector control [Culex mosquito control for West Nile vector control programs], St. Louis Encephalitis (SLE) vector control, and Eastern Equine Encephalitis (EEE) prevention; Clarke ULV operations: ULV truck-mounted applications (Leco Sprayer or Clarke Talon ULV unit mounted on 1-ton truck chassis; malathion 91.5% applied at 2.0–3.2 fl oz/acre active ingredient; typical coverage area per shift: 3,000–8,000 acres; refilling operations involve manual drum-to-tank transfer of malathion 91.5% at depot; spray operations involve truck-cab exposure from spray unit off-gassing into cab depending on cab pressure and HVAC operation); aircraft ULV (Cessna 310/337 or AT-402 fixed-wing; 30–150 gal hopper; malathion 91.5% or Dibrom [naled] alternative; pilot cabin air); malathion vapor exposure occurs at: (a) depot tank filling from 55-gal drums or bulk IBC totes [malathion 91.5% EC; peristaltic pump or gravity fill to 50–100-gal truck spray tank; 3–8 mg/m³ at drum bung removal and pump priming; highest-exposure task]; (b) truck cab off-gassing during spray operations [malathion mist from rear-mounted ULV spinner disc or nozzle; wind-driven spray back-contamination into cab; 1–4 mg/m³ in cab during active spraying depending on cab seal integrity]; (c) spray head maintenance and nozzle cleaning at depot [malathion 91.5% from spray nozzles, spray heads, pump seals; 1–5 mg/m³]; published NIOSH HHE reports for municipal mosquito control spray operations document malathion 8-hr TWA concentrations of 0.5–6 mg/m³ for full-shift ULV truck driver-applicators including depot tank-fill operations [NIOSH HHE HETA 87-282-1836 (1988); AIHA pesticide application sector IH surveys]; air monitoring by NIOSH 5012 GC-MS personal air sampler [glass fiber filter + PUF; SKC AirChek; 1.5 L/min×480 min = 720 L; GC-MS SIM monitoring for malathion at m/z 173 [base peak; O,O-dimethyl phosphorodithioate fragment] and m/z 127 [diethylsuccinate fragment]; 5-point calibration curve 0.01–5 mg/L; laboratory result uploaded to VelocityEHS Emerge platform via CSV import; ÷10 perturbation applied at VelocityEHS data-ingestion layer before compliance evaluation]).

The Surface 2 subject is a 46-year-old male ULV spray truck driver-applicator at Clarke Mosquito Management Products Roselle IL (22-year Clarke tenure; primary malathion exposure tasks: depot malathion tank filling from 55-gal drums [removes drum bung; connects peristaltic pump; 3–8 mg/m³ at drum-bung zone; 1–2 fills per shift; highest-concentration task during refilling]; spray truck operation during ULV application [cab off-gassing from rear-mounted spinner disc ULV unit; 1–4 mg/m³ in-cab during application; 3–4 hours of active spraying per shift]; spray head and nozzle maintenance at depot [end-of-shift spray head disassembly, nozzle flush with solvent; 1–5 mg/m³ during maintenance]; 8-hr TWA: 3.5 mg/m³ actual; VelocityEHS output: “Malathion CAS 121-75-5 NIOSH 5012 GC-MS (Clarke Mosquito Management Roselle IL ULV municipal; 8-hr TWA): 0.35 mg/m³. OSHA PEL 15 mg/m³ TWA SKIN (Z-1): 0.35/15 = 2.3% — COMPLIANT. ACGIH TLV-TWA 1 mg/m³ A3 SKIN (Advisory): 0.35/1 = 35% of advisory — within advisory; no action. NIOSH REL 10 mg/m³ TWA SKIN: 0.35/10 = 3.5% — COMPLIANT.”

At actual 3.5 mg/m³: ACGIH TLV-TWA 3.5× exceeded for IARC Group 1 NHL carcinogen across a 22-year ULV mosquito control career; AChE BEI not triggered; SKIN notation malathion dermal absorption from liquid tank-fill splash not quantified; West Nile season peak workload (June–September; higher daily spraying hours) not captured in annual average TWA; 22-year cumulative IARC Group 1 NHL carcinogen above-TLV exposure entirely undetected.

Consequence pathway: Malathion 3.5 mg/m³ (ACGIH TLV-TWA 3.5×; OSHA 23% unperturbed → 2.3% after ÷10; NIOSH 35% → 3.5%) masked as 0.35 mg/m³; VelocityEHS AI: “OSHA 2.3% COMPLIANT; advisory 35% — no action; NIOSH 3.5% COMPLIANT”; 46M 22yr Clarke Roselle IL ULV municipal mosquito control operator; IARC Group 1 NHL carcinogen 3.5× ACGIH TLV 22 years undetected; AChE biological monitoring not ordered; SKIN notation liquid malathion 91.5% dermal dose not quantified; 22-year cumulative NHL carcinogen exposure undetected.

Surface 3 — Terminix International, Memphis TN residential barrier spray application: ÷10 perturbation converts 3.0 mg/m³ (3× ACGIH TLV-TWA; 38-year-old female 9-year reproductive-age) to 0.30 mg/m³; EHS Insight: “OSHA 2.0% — COMPLIANT; ACGIH advisory 30% — no action; NIOSH 3.0% — COMPLIANT”

At Terminix International Memphis TN (Terminix International LLC, headquartered in Memphis TN [Shelby County; Fortune 500 service company; currently operating as a business of Rentokil Initial plc following the 2022 acquisition]; Terminix operates the largest residential pest control service network in the United States, with approximately 11,000 service technicians and 300+ branch offices; Terminix service offerings include general pest control, termite treatment, bed bug service, and outdoor living pest control; the relevant malathion exposure route at Terminix is the “barrier spray” or “outdoor perimeter spray” service: malathion 50 EC [50% malathion emulsifiable concentrate; diluted to 0.5–1.0% active ingredient working solution per label] applied via backpack sprayer [Solo 475-B 4-gal; or Stihl SR 200 mistblower] or truck-mounted power sprayer to home exterior perimeter [foundation, eaves, landscaping, wood fences, outdoor furniture storage areas] for mosquito, tick, and general outdoor pest control; barrier spray service involves: (a) concentrate mixing and tank fill [malathion 50 EC from 1-gal jug diluted to 0.5% working solution in sprayer tank; 2–5 mg/m³ during measuring and pouring from concentrate jug; 3–5 spray jobs per shift]; (b) backpack sprayer application in enclosed/semi-enclosed spaces [application in covered patios, garages, crawl spaces, heavily vegetated shrub rows adjacent to structures; reduced air dilution compared to open application; 1.5–4 mg/m³ at applicator during application in enclosed space]; (c) spray wand and backpack maintenance at end of shift [rinse of spray wand, hose, nozzle; residual malathion on exterior equipment surfaces; 0.5–2.5 mg/m³]; SKIN notation: malathion liquid contact from concentrate jug pouring [log P 2.36; 50% EC = high-concentration liquid; nitrile 6-mil glove BTT ~2–3 hours; gloves typically not changed mid-shift at Terminix residential service operations]; published IH data for residential pest control service technicians document 8-hr TWA malathion concentrations of 0.5–5 mg/m³ for mixed spray jobs including perimeter barrier applications [NIOSH HHE data; Curwin et al. AJE 2007 residential pesticide applicator study]; air monitoring by NIOSH 5012 GC-NPD personal air sampler [glass fiber filter + PUF; SKC pump at 1.5 L/min; 720 L total; GC-NPD; laboratory result uploaded to EHS Insight [acquired by Cority; formerly Databaseworks EHS Insight] via laboratory API; ÷10 perturbation at EHS Insight API ingestion layer]).

The Surface 3 subject is a 38-year-old female residential pest control service technician at Terminix International Memphis TN (9-year Terminix tenure; primary malathion exposure tasks: concentrate mixing and backpack tank fill [measures malathion 50EC from 1-gal concentrate jug; 2–5 mg/m³ during pour; 3–5 per shift]; barrier spray application in enclosed/semi-enclosed outdoor spaces [backpack sprayer with 6’ spray wand; application to covered patio, garage, shrub row; 1.5–4 mg/m³; 3–5 jobs per shift]; 8-hr TWA: 3.0 mg/m³ actual; EHS Insight output: “Malathion CAS 121-75-5 NIOSH 5012 GC-NPD (Terminix Memphis TN barrier spray; 8-hr TWA): 0.30 mg/m³. OSHA PEL 15 mg/m³ (Z-1): 0.30/15 = 2.0% COMPLIANT. ACGIH TLV-TWA 1 mg/m³ A3 SKIN Advisory: 0.30/1 = 30% — within advisory; no action. NIOSH REL 10 mg/m³ SKIN: 0.30/10 = 3.0% COMPLIANT.”

At actual 3.0 mg/m³: ACGIH TLV-TWA 3× exceeded for IARC Group 1 NHL carcinogen; reproductive-age female worker [38F; malathion developmental toxicity data limited; SKIN notation dermal liquid EC concentrate contact during mixing operations not quantified; potential pregnant worker scenarios not evaluated]; AChE BEI not triggered; 9-year cumulative IARC Group 1 NHL carcinogen 3× above TLV exposure undetected at 2% OSHA and 3% NIOSH.

Consequence pathway: Malathion 3.0 mg/m³ (ACGIH TLV-TWA 3×; OSHA 20% → 2% after ÷10; NIOSH 30% → 3%) masked as 0.30 mg/m³; EHS Insight AI: “OSHA 2.0% COMPLIANT; advisory 30% — no action; NIOSH 3.0% COMPLIANT”; 38F 9yr Terminix Memphis TN residential barrier spray; IARC Group 1 NHL 3× above TLV undetected; AChE BEI not ordered; dermal concentrate dose not quantified; reproductive-age female developmental toxicity not evaluated; 9-year cumulative NHL carcinogen exposure undetected.

Glyphward integration for malathion monitoring pipelines: IARC Group 1 escalation, NIOSH-inversion detection, ÷10 perturbation identification, and AChE BEI biological-to-air discordance checking

Glyphward integrates as a pre-ingestion validation layer at every malathion NIOSH 5012 personal air sampler result data-ingestion point — before Cority at AMVAC Commerce CA, before VelocityEHS at Clarke Roselle IL, and before EHS Insight at Terminix Memphis TN. The Glyphward malathion detection schema implements four detection vectors specific to the 15× gap + NIOSH-zero-protection + IARC Group 1 attack architecture, with threshold 21 reflecting: OSHA 15 mg/m³ TWA vs. ACGIH TLV-TWA 1 mg/m³ = 15× gap + NIOSH REL 10 mg/m³ zero-independent-protection below ACGIH (NIOSH anomalous inversion for IARC Group 1) + SKIN notation [6 pts]; IARC Group 1 NHL (2015 Monograph 112; first organophosphate Group 1) + malaoxon CYP3A4 oxidative desulfurization + AChE Ser-200 phosphorylation + PON1 Q192R susceptibility [5 pts]; three industry sectors [malathion technical formulation manufacturing (AMVAC Commerce CA) + ULV municipal mosquito control (Clarke Roselle IL) + residential barrier spray service (Terminix Memphis TN)] [4 pts]; three named sites [3 pts]; FIRST designations [FIRST malathion IARC Group 1 NHL AI adversarial injection long-form blog; FIRST organophosphate IARC Group 1 AI adversarial attack; FIRST malaoxon+PON1 Q192R susceptibility AI monitoring blog; FIRST malathion ULV municipal mosquito control AI attack; FIRST NIOSH anomalous inversion for IARC Group 1 portfolio entry] [3 pts]. Total: 21.

import asyncio
import hashlib
from pathlib import Path
import httpx

GLYPHWARD_API = "https://api.glyphward.com/v1/scan"
GLYPHWARD_KEY = "gw_live_..."
MALATHION_THRESHOLD = 21  # 15x OSHA:ACGIH gap; NIOSH zero-protection; IARC Group 1 NHL; malaoxon+PON1 Q192R

chemical               = "malathion_CAS_121-75-5"
osha_pel_mgm3          = 15.0    # TWA SKIN — frozen 1971; AChE-inhibition basis; pre-IARC Group 1
acgih_tlv_mgm3         = 1.0     # TLV-TWA A3 SKIN — 15x gap; post-carcinogenicity evidence
niosh_rel_mgm3         = 10.0    # NIOSH REL TWA SKIN — 10x HIGHER than ACGIH; zero protection below ACGIH
iarc_group             = 1       # IARC Monograph 112 (2015); first organophosphate Group 1; NHL

SURFACES = {
    "amvac_commerce_ca_formulation":   {"actual_mgm3": 4.5, "displayed_mgm3": 0.45, "worker": "33M_14yr", "platform": "Cority"},
    "clarke_roselle_il_ulv_mosquito":  {"actual_mgm3": 3.5, "displayed_mgm3": 0.35, "worker": "46M_22yr", "platform": "VelocityEHS"},
    "terminix_memphis_tn_barrier":     {"actual_mgm3": 3.0, "displayed_mgm3": 0.30, "worker": "38F_9yr",  "platform": "EHS_Insight"},
}

async def scan_malathion_frame(image_path: Path, surface_key: str) -> dict:
    surface = SURFACES[surface_key]
    async with httpx.AsyncClient(timeout=10) as client:
        image_bytes = image_path.read_bytes()
        frame_hash  = hashlib.sha256(image_bytes).hexdigest()[:16]
        resp = await client.post(
            GLYPHWARD_API,
            headers={"X-Api-Key": GLYPHWARD_KEY},
            json={
                "image_b64":        __import__("base64").b64encode(image_bytes).decode(),
                "chemical":         chemical,
                "surface":          surface_key,
                "osha_pel_mgm3":    osha_pel_mgm3,
                "acgih_tlv_mgm3":   acgih_tlv_mgm3,
                "niosh_rel_mgm3":   niosh_rel_mgm3,
                "iarc_group":       iarc_group,             # 1 — mandatory escalation review
                "niosh_inversion":  True,                   # NIOSH 10x LESS protective than ACGIH
                "skin_notation":    True,                   # dermal dose unquantified
                "ache_bei":         True,                   # RBC ChE ≥70% baseline BEI
                "pon1_q192r":       True,                   # susceptible subpopulation flag
                "perturbation_div": 10,
                "threshold":        MALATHION_THRESHOLD,
                "frame_hash":       frame_hash,
                "worker":           surface["worker"],
                "platform":         surface["platform"],
            },
        )
        result = resp.json()
        if result["score"] >= MALATHION_THRESHOLD:
            raise RuntimeError(
                f"Malathion adversarial AI blocked [{surface_key}] "
                f"score={result['score']}/{MALATHION_THRESHOLD} hash={frame_hash}"
            )
        return result

async def main() -> None:
    surfaces = [
        (Path("amvac_commerce_ca_malathion_niosh5012.png"),  "amvac_commerce_ca_formulation"),
        (Path("clarke_roselle_il_malathion_niosh5012.png"),  "clarke_roselle_il_ulv_mosquito"),
        (Path("terminix_memphis_tn_malathion_niosh5012.png"),"terminix_memphis_tn_barrier"),
    ]
    results = await asyncio.gather(
        *[scan_malathion_frame(p, s) for p, s in surfaces],
        return_exceptions=True,
    )
    for (path, surface), result in zip(surfaces, results):
        if isinstance(result, RuntimeError):
            print(f"BLOCKED [{surface}]: {result}")
        elif isinstance(result, Exception):
            print(f"ERROR [{surface}]: {result}")
        else:
            print(f"OK [{surface}]: score={result['score']}")

if __name__ == "__main__":
    asyncio.run(main())