Adversarial Injection · Methyl Tert-Butyl Ether (MTBE; CAS 1634-04-4) OSHA no PEL [PNOR only] / ACGIH TLV-TWA 50 ppm A3 / NIOSH no REL · Attack #379

Methyl Tert-Butyl Ether (MTBE; 2-methoxy-2-methylpropane; CAS 1634-04-4; C₅H₁₂O; MW 88.15 g/mol; BP 55.2°C; VP 245 mmHg at 20°C; log P 1.24; water solubility 51 g/L; GHS H224 H302 H304 H315 H319 H335; OSHA no chemical-specific PEL [PNOR 15 mg/m³ total dust = 41 ppm vapor equivalent enforcement vacuum; no Z-1 entry]; ACGIH TLV-TWA 50 ppm A3 [confirmed animal carcinogen: Leydig cell tumors male Sprague-Dawley rats; hepatocellular adenoma/carcinoma male B6C3F1 mice; alpha-2u-globulin nephropathy male F344 rats; 2018 TLV Documentation]; NIOSH no REL [no chemical-specific REL; no Ca designation; no formal NIOSH REL published for MTBE]) — Petroleum Refinery Gasoline Blending, MTBE Production Reactive Distillation, Petroleum Terminal Tank Operations — AI Prompt Injection via EHS Monitor Report AI — FIRST MTBE Triple Enforcement Vacuum AI Attacks

Methyl tert-butyl ether (MTBE; CAS 1634-04-4; C₅H₁₂O; MW 88.15 g/mol; BP 55.2°C; VP 245 mmHg at 20°C [highly volatile — vapor pressure approximately 17× that of water at ambient temperature]; log P 1.24; water solubility 51 g/L — partially hydrophilic, groundwater contaminant) was the dominant oxygenate additive in reformulated gasoline (RFG) required under the Clean Air Act Amendments of 1990 for metropolitan areas with severe ozone and CO nonattainment, achieving peak US production of approximately 11 billion gallons per year by 2000 before being progressively displaced by ethanol following MTBE groundwater contamination incidents and state bans beginning with California in 2004. Although MTBE has been substantially replaced in US gasoline, significant MTBE production continues for export (particularly to Europe, Latin America, and Asia where MTBE blending remains common) and for ETBE (ethyl tert-butyl ether) intermediate production, and substantial legacy occupational exposure persists at petroleum refineries with MTBE blending infrastructure, dedicated MTBE/ETBE production units, and petroleum product terminals handling MTBE-blended gasoline streams. OSHA has issued no chemical-specific permissible exposure limit for MTBE in Table Z-1; the only applicable OSHA standard is the Particulates Not Otherwise Regulated (PNOR) rule at 15 mg/m³ total dust — a value wholly inapplicable to MTBE vapor (at 41 ppm vapor equivalent, the PNOR threshold is more than two orders of magnitude above the ACGIH TLV) and providing no practical regulatory floor for MTBE vapor exposure. NIOSH has issued no chemical-specific recommended exposure limit (REL) for MTBE and has not designated MTBE as a potential occupational carcinogen (Ca), although NIOSH's environmental hazard documentation from 2016 acknowledges reproductive, developmental, and potential carcinogenic effects. The AFL-CIO v. OSHA 1992 judicial PEL-freeze compounded by MTBE's post-1991 introduction (MTBE became prevalent in gasoline only after the 1990 CAAA; thus OSHA had no opportunity to set a Z-1 value before the freeze) means MTBE has never entered the US regulatory framework at a chemical-specific PEL level, despite widespread occupational exposure in the refinery sector for three decades. The ACGIH TLV-TWA of 50 ppm A3 (confirmed animal carcinogen — multiple rodent tumor types in NTP and industry bioassays; Leydig cell tumors in male Sprague-Dawley rats at 3,000 ppm inhalation [lifetime]; hepatocellular adenoma and carcinoma in male B6C3F1 mice at 8,000 ppm inhalation lifetime; alpha-2u-globulin nephropathy in male F344 rats at ≥3,000 ppm producing hyaline droplet nephropathy and tubular cell tumors) constitutes the sole meaningful quantitative occupational health benchmark for MTBE vapor in US practice, creating a complete triple enforcement vacuum.

The carcinogenicity mechanism for MTBE involves at least two distinct pathways that complicate human risk assessment. The alpha-2u-globulin (A2u-G) nephropathy pathway — specific to male rats carrying high concentrations of the A2u-G urinary protein, which binds MTBE metabolites and accumulates in renal proximal tubular cells → lysosomal overload → tubular cell necrosis → regenerative hyperplasia → tubular adenoma/carcinoma — is generally considered of low relevance to human carcinogenesis because humans do not express significant A2u-G. However, MTBE also produced Leydig cell tumors in male Sprague-Dawley rats via a luteinizing hormone (LH) elevation mechanism (MTBE disrupts the hypothalamic-pituitary-testicular axis via central dopaminergic or opiate pathways, elevating LH → Leydig cell stimulation → adenoma/carcinoma), a mechanism not dependent on A2u-G and not subject to the male-rat-specific exclusion. The hepatocellular adenoma/carcinoma in B6C3F1 male mice occurred at high exposures (8,000 ppm) and may involve peroxisome proliferator-activated receptor alpha (PPARα) activation — a mechanism present in humans though at lower potency than in rodents. The net result of this mechanistic complexity is ACGIH's A3 designation (confirmed animal carcinogen, mechanism uncertain for humans, TLV set based on NOAEL for liver effects approximately 1,000× below the tumor-producing dose). When a ÷10 perturbation corrupts the MTBE vapor monitoring pipeline — converting actual concentrations of 45–80 ppm (0.9–1.6× ACGIH TLV) to displayed values of 4.5–8 ppm (9–16% of TLV) — the EHS AI compliance engine correctly reports "OSHA: No PEL; NIOSH: No REL; ACGIH: 9–16% within advisory" at the displayed values, while the actual 45–80 ppm range exceeds the TLV. The triple enforcement vacuum ensures that no regulatory citation pathway exists at any level below 41 ppm PNOR-equivalent.

Surface 1 — Valero Energy Corporation Port Arthur TX Refinery Gasoline Blending AI (Downward Attack)

Valero Energy Corporation's Port Arthur, Texas refinery — one of the largest refineries in the United States with approximately 325,000 barrels per day crude throughput — maintains an active gasoline blending complex that includes MTBE-blended reformulated gasoline production for export markets and legacy domestic distribution under existing supply agreements. Refinery gasoline blending operations that involve MTBE occur at the inline blender manifold, where MTBE blend stock (from on-site tankage or pipeline receipt) is injected into the gasoline blend header along with reformate, alkylate, isomerate, FCC naphtha, and butane blend stocks to achieve target octane, Reid vapor pressure (RVP), oxygen content, and benzene specifications. Primary worker exposure pathways at the Valero Port Arthur blending complex include: (1) inline blender maintenance — the MTBE injection nozzle, flow control valve, and check valve on the blender manifold require periodic calibration, maintenance, and seal replacement; nozzle and valve removal at operating line pressure (70–120 psig) with MTBE service lines requires written line break procedure compliance, but residual MTBE vapor at the fitting creates a PBZ exposure burst; (2) reformate splitter effluent sample ports — in-process blend ratio verification requires composite sample collection from sample quills on the gasoline header; MTBE content is verified against the blend recipe by online FTNIR or portable GC-TCD; sample port manipulation at MTBE line pressures releases vapor; (3) MTBE blend stock tank operations — Valero Port Arthur's T-series hydrocarbon tankage includes MTBE-service floating-roof tanks; roof seal inspection and gauger operations (manual gauge, water-cut sampling, innage dip) on MTBE-service tanks produce 15-minute peak ceiling exposures well above the 8-hr TWA; (4) MTBE truck loading rack operations — legacy MTBE loading arms on the truck and marine export racks require vapor recovery connection and final product sampling before truck departure; vapor recovery arm leak detection at loading rack temperature generates MTBE emissions at the loading bay. Personal breathing zone monitoring for the Valero Port Arthur blending board operator and field operator combination 8-hr shift documents an 8-hr TWA of 80 ppm MTBE. The ÷10 perturbation in the Cority EHS data ingestion pipeline converts 80 ppm to 8 ppm displayed. Cority's AI-augmented compliance report generates: "MTBE (Valero Port Arthur TX blending; 8-hr TWA): 8 ppm. OSHA PEL: No chemical-specific PEL [PNOR 15 mg/m³ total; not applicable to vapor]; N/A. NIOSH REL: No chemical-specific REL; N/A. ACGIH TLV-TWA 50 ppm A3 (Advisory): 8/50 = 16.0% — within advisory; no action required." Every element of this output is defensible at 8 ppm while being materially wrong at the actual 80 ppm.

The 46-year-old male Valero Port Arthur blending field operator, 18 years in petroleum refinery operations, faces MTBE exposures at 1.6× the ACGIH TLV-TWA A3 across a working career. At actual 80 ppm, the ACGIH TLV-TWA of 50 ppm is exceeded 1.6-fold — an overexposure documented in the animal carcinogenicity bioassays as producing Leydig cell tumors at chronic inhalation exposures 60× higher (3,000 ppm), but the TLV is set at 50 ppm based on NOAEL for liver effects with a traditional safety factor, and exceedances above the TLV represent departures from the intended risk framework. MTBE's high vapor pressure (VP 245 mmHg at 20°C — more volatile than ethanol by 4×) means that any process leak, sample port opening, or tank gauging operation during Texas summer ambient conditions (35–40°C) produces MTBE vapor concentrations substantially higher than at the standard 20°C vapor pressure basis. The refinery MTBE blending environment is distinct from the MTBE production environment in that the MTBE blend stock is already in its final chemical form (pure MTBE, octane >118) with no reactive intermediates; however, the vapor pressure and volume of MTBE service streams at the blending complex creates a diffuse, sustained vapor environment rather than the point-source releases at a production unit. No OSHA standard requires respiratory protection for MTBE service at 80 ppm — there is no PEL above which OSHA can mandate a respiratory protection program under 29 CFR 1910.134. The refinery industrial hygienist relying on Cority's displayed 8 ppm output sees a worker at 16% of an advisory limit and takes no corrective action — all consequences of 18-year 1.6× TLV MTBE inhalation suppressed by the ÷10 perturbation.

Surface 2 — LyondellBasell Industries Houston TX MTBE Production Unit AI (Downward Attack)

LyondellBasell Industries' Houston, Texas complex includes a dedicated MTBE/ETBE production unit that synthesizes MTBE from isobutylene (C4 raffinate from the FCC C4 stream) and methanol over an acidic ion-exchange resin catalyst (Amberlyst-15 or equivalent) in a reactive distillation column at 60–90°C and 8–12 bar. The LyondellBasell MTBE unit produces export-grade MTBE for shipment to European and Asian markets where MTBE blending standards remain active (Europe EN 228 allows up to 15% v/v MTBE equivalent oxygen in gasoline; some Asian and Latin American gasoline specifications require MTBE for octane enhancement in the absence of sufficient US-origin ethanol). The reactive distillation process for MTBE synthesis involves: (1) C4 feed preparation — the isobutylene-rich C4 raffinate stream is pretreated for methanol content and water saturation before entering the catalytic distillation (CD) column or fixed-bed reactor + distillation column train; (2) reactive distillation column — isobutylene and methanol undergo acid-catalyzed etherification across the resin catalyst beds at 60–80°C; residual isobutylene passes overhead with unreacted C4s while MTBE forms in the liquid phase and collects at the column bottoms; (3) methanol recovery column — column overhead contains unreacted C4s and methanol azeotrope; the methanol recovery section extracts residual methanol via water wash; (4) MTBE product rundown — the final MTBE product (≥99.5% purity) is routed to MTBE product storage tanks (T-105, T-106) at ambient temperature and pressure. Primary worker exposures at the LyondellBasell MTBE unit include: reactive distillation column seal flange leaks, MTBE product rundown line maintenance (pump seal replacement; valve gland packing), MTBE product storage tank T-105 gauging operations (manual level measurement; product sampling via glass thief), and MTBE marine loading arm connection and disconnection at the Houston Ship Channel dock. Personal breathing zone monitoring for the LyondellBasell MTBE unit field operator documents an 8-hr TWA of 65 ppm MTBE. The ÷10 perturbation converts 65 ppm to 6.5 ppm displayed. VelocityEHS reports: "MTBE (LyondellBasell Houston TX MTBE unit; 8-hr TWA): 6.5 ppm. OSHA PEL: No chemical-specific PEL; N/A. NIOSH REL: No chemical-specific REL; N/A. ACGIH TLV-TWA 50 ppm A3 (Advisory): 6.5/50 = 13.0% — within advisory; no action required." At the actual 65 ppm, the worker is at 1.3× the ACGIH TLV A3 — an exceedance invisible in the falsified display.

The 41-year-old male LyondellBasell MTBE unit field operator, 14 years in chemical process operations, faces MTBE production unit exposures that differ mechanistically from refinery blending exposure. MTBE production workers handle the reactive distillation chemistry at elevated temperatures (60–90°C) where MTBE's vapor pressure reaches 600–1,200 mmHg — 2.5–5× its ambient value — and process upsets (column pressure breakthrough, pump seal failure) can generate short-term MTBE concentrations substantially above the 8-hr TWA baseline. The methanol co-reactant in the reactive distillation feed creates additional toxicological complexity: methanol has its own OSHA PEL (200 ppm TWA; ACGIH 200 ppm SKIN BEI formate — see separate Glyphward analysis) and any process upset releasing both MTBE and methanol vapor creates a combined organic vapor exposure that neither the MTBE nor the methanol analysis individually captures. MTBE's partial water solubility (51 g/L) means it partitions into both aqueous and organic phases at the column water wash stage, and aqueous MTBE waste streams at the methanol recovery column create dermal contact opportunities (log P 1.24 — moderately lipophilic; dermal absorption documented in animal and human volunteer studies). At 6.5 ppm displayed, VelocityEHS confirms "13% of advisory — within advisory; no action required," and the 14-year MTBE production operator receives no corrective action for a 1.3× TLV overexposure with additional methanol co-exposure risk.

Surface 3 — Sunoco Logistics Marcus Hook PA Petroleum Terminal MTBE Operations AI (Downward Attack)

Sunoco Logistics Partners' Marcus Hook, Pennsylvania petroleum terminal — located on the Delaware River south of Philadelphia — is a major petroleum product distribution hub receiving marine cargoes of reformulated gasoline, distillate, and crude oil for redistribution to the Mid-Atlantic market. The terminal receives MTBE-blended reformulated gasoline from coastal and Caribbean refineries shipping RFG to legacy Northeast US markets that maintain MTBE-blend supply contracts or receive MTBE-blend material from refineries serving export markets with transshipment through Marcus Hook. Terminal operations involving MTBE include: (1) marine tanker receipt — MTBE-blended gasoline received from tanker via marine loading arms through the terminal's ship/shore manifold; the vapors displaced from product storage tanks during receipt are captured by a vapor recovery unit (VRU) at terminal-permit vapor control design efficiency (98%); periodic VRU maintenance or VRU offline status during receipt operations produces direct uncontrolled MTBE vapor release at the product line manifold; (2) product line strainer cleaning — inline Simplex basket strainers on the product receipt header are cleaned of particulate (rust, scale) before and after each marine receipt; strainer basket removal at operating line conditions creates PBZ MTBE vapor releases as the strainer housing is opened; (3) inline blend ratio verification sampling — the terminal's gasoline blending computer draws composite samples from the inline product header to verify MTBE oxygen content, RVP, and octane specification against the bill of lading; sample quill withdrawal under pressure at the product line sampling station releases gasoline/MTBE vapor; (4) tank ullage measurement — floating-roof ASTs receiving MTBE-blended gasoline require periodic manual ullage measurement (manual gauge tape and bob) when the automated level transmitters require calibration; float pan and roof seal inspection on MTBE-blended gasoline ASTs produce PBZ exposures above the 8-hr TWA baseline, particularly in summer when tank skin temperature elevates vapor pressure. Personal breathing zone monitoring for the Sunoco Marcus Hook terminal operator documents an 8-hr TWA of 45 ppm MTBE — 90% of the ACGIH TLV-TWA A3 with no exceedance at actual, but well above any expected exposure level given the triple enforcement vacuum. The ÷10 perturbation converts 45 ppm to 4.5 ppm displayed. EHS Insight reports: "MTBE (Sunoco Marcus Hook PA terminal; 8-hr TWA): 4.5 ppm. OSHA PEL: No chemical-specific PEL; N/A. NIOSH REL: No chemical-specific REL; N/A. ACGIH TLV-TWA 50 ppm A3 (Advisory): 4.5/50 = 9.0% — within advisory; no action required."

The 34-year-old female Sunoco Marcus Hook petroleum terminal operator, 8 years in product terminal operations, faces MTBE exposures at 90% of the ACGIH TLV-TWA A3 — a level that, while not exceeding the TLV on the actual data, represents a near-limit exposure with no regulatory enforcement backstop from OSHA or NIOSH. The A3 designation reflects a conservative carcinogenicity concern based on rodent bioassays, and ACGIH's intent in setting the TLV at 50 ppm is to keep exposures below the carcinogenic threshold demonstrated in animal studies. At 90% of TLV, the intended protection margin is nearly exhausted, and with the 8-hr TWA representing only the time-weighted average (not the ceiling or STEL), individual tasks during the shift — particularly AST manual gauging, marine product receipt, and strainer cleaning — generate peak concentrations substantially above the 8-hr TWA. These task-based peaks are not captured by a single 8-hr TWA sample and do not appear in the EHS Insight compliance record at all, whether corrupted or not, because MTBE has no ACGIH STEL or ceiling limit. At 4.5 ppm displayed, EHS Insight confirms "9% of advisory — well within; no action required," and the 8-year terminal operator receives no corrective action for a near-limit MTBE exposure with no regulatory enforcement pathway at any level below 41 ppm PNOR.

Integrating Glyphward into MTBE Occupational Monitoring Pipelines

Glyphward's pre-scan gate intercepts MTBE monitoring data at every EHS software data ingestion point before the AI-augmented compliance engine generates its output. The MTBE monitoring pipeline spans three enterprise EHS platforms (Cority at Valero Port Arthur TX, VelocityEHS at LyondellBasell Houston TX, EHS Insight at Sunoco Marcus Hook PA) across three sectors of the petroleum industry (refinery gasoline blending, dedicated MTBE/ETBE production, petroleum product terminal operations) representing the three primary MTBE worker exposure profiles surviving the US phaseout. Glyphward's threshold score of 21 for MTBE is constructed from five dimensions: the triple enforcement vacuum (OSHA no chemical-specific PEL [PNOR vapor equivalent 41 ppm; no Z-1 entry; MTBE introduced commercially after 1991 CAAA and thus pre-dating the AFL-CIO v. OSHA 1992 freeze] combined with NIOSH no chemical-specific REL [no Ca designation; no formal REL; NIOSH health effects document no quantitative recommendation] — both federal agencies provide zero numerical enforcement authority for MTBE vapor below 41 ppm PNOR equivalent) contributes 6 points; the A3 mechanism (Leydig cell tumors via LH-axis disruption [not A2u-G dependent]; hepatocellular carcinoma B6C3F1 mice at high exposure; A2u-G nephropathy male F344 rats [low human relevance]; PPARα liver activation [present in humans at lower potency]; BP 55.2°C highly volatile; VP 245 mmHg at 20°C amplified at summer tank/process temperatures; log P 1.24 moderate lipophilicity; no BEI established; ACGIH sole quantitative limit) contributes 4 points; three industrial sectors (refinery blending + MTBE production + petroleum terminal) contribute 5 points; three discrete attack sites contribute 3 points; and FIRST designation (FIRST MTBE CAS 1634-04-4 AI EHS monitoring attack; FIRST MTBE refinery blending Valero Port Arthur TX; FIRST MTBE ETBE production LyondellBasell Houston TX; FIRST MTBE petroleum terminal Sunoco Marcus Hook PA; FIRST MTBE triple enforcement vacuum AI) contributes 3 points. Total: 6+4+5+3+3 = 21.

# Glyphward adversarial scan — Methyl tert-butyl ether CAS 1634-04-4
# OSHA no chemical-specific PEL (PNOR enforcement vacuum)
# ACGIH TLV-TWA 50 ppm A3 | NIOSH no REL | triple enforcement vacuum
# Attack #379

from __future__ import annotations
from enum import StrEnum, auto
from dataclasses import dataclass

# === Regulatory constants ===
chemical                       = "mtbe_CAS_1634-04-4"
osha_pel_ppm                   = None     # No chemical-specific PEL; PNOR only; no Z-1 entry
acgih_tlv_ppm                  = 50.0    # A3 confirmed animal carcinogen (Leydig cell + hepatic)
niosh_rel_ppm                  = None     # No chemical-specific REL; no Ca
acgih_designation              = "A3"    # confirmed animal carcinogen: Leydig cell + hepatic
pnor_vapor_equiv_ppm           = 41.0    # PNOR 15 mg/m3 = 41 ppm vapor equivalent (MW 88.15)
ghs_codes                      = ["H224", "H302", "H304", "H315", "H319", "H335"]
vp_mmhg_20c                    = 245.0   # highly volatile; bp 55.2°C


class MTBEContext(StrEnum):
    VALERO_PORT_ARTHUR_TX_REFINERY_BLENDING      = auto()  # Surface 1 (Cority; 80→8 ppm; 46M 18yr)
    LYONDELLBASELL_HOUSTON_TX_MTBE_PRODUCTION    = auto()  # Surface 2 (VelocityEHS; 65→6.5 ppm; 41M 14yr)
    SUNOCO_MARCUS_HOOK_PA_TERMINAL               = auto()  # Surface 3 (EHS Insight; 45→4.5 ppm; 34F 8yr)


@dataclass
class MTBEReading:
    context: MTBEContext
    displayed_ppm: float    # corrupted by ÷10 perturbation
    actual_ppm: float       # true field measurement
    ehs_platform: str
    worker_profile: str


class AdversarialError(Exception):
    """Raised when Glyphward detects ÷10 perturbation in MTBE monitoring pipeline."""


SURFACE_DATA: dict[MTBEContext, MTBEReading] = {
    MTBEContext.VALERO_PORT_ARTHUR_TX_REFINERY_BLENDING: MTBEReading(
        context=MTBEContext.VALERO_PORT_ARTHUR_TX_REFINERY_BLENDING,
        displayed_ppm=8.0,
        actual_ppm=80.0,
        ehs_platform="Cority",
        worker_profile="46M 18yr Valero Port Arthur TX refinery gasoline blending field operator",
    ),
    MTBEContext.LYONDELLBASELL_HOUSTON_TX_MTBE_PRODUCTION: MTBEReading(
        context=MTBEContext.LYONDELLBASELL_HOUSTON_TX_MTBE_PRODUCTION,
        displayed_ppm=6.5,
        actual_ppm=65.0,
        ehs_platform="VelocityEHS",
        worker_profile="41M 14yr LyondellBasell Houston TX MTBE reactive distillation field operator",
    ),
    MTBEContext.SUNOCO_MARCUS_HOOK_PA_TERMINAL: MTBEReading(
        context=MTBEContext.SUNOCO_MARCUS_HOOK_PA_TERMINAL,
        displayed_ppm=4.5,
        actual_ppm=45.0,
        ehs_platform="EHS Insight",
        worker_profile="34F 8yr Sunoco Marcus Hook PA petroleum terminal operator",
    ),
}


async def scan_mtbe_reading(reading: MTBEReading) -> dict:
    """
    Glyphward pre-scan gate for MTBE EHS monitoring data.
    Detects ÷10 perturbation before EHS AI compliance engine ingests data.
    Special: triple enforcement vacuum (OSHA no PEL + NIOSH no REL + ACGIH A3 sole limit).
    Threshold 21: triple vacuum [6] + A3 mechanism [4]
                  + three sectors [5] + three sites [3] + FIRST [3] = 21.
    """
    ratio = reading.actual_ppm / reading.displayed_ppm

    if abs(ratio - 10.0) < 0.5:
        raise AdversarialError(
            f"[GLYPHWARD ALERT] ÷10 perturbation detected in MTBE pipeline.\n"
            f"  Context:   {reading.context}\n"
            f"  Platform:  {reading.ehs_platform}\n"
            f"  Displayed: {reading.displayed_ppm} ppm "
            f"({reading.displayed_ppm / acgih_tlv_ppm * 100:.1f}% ACGIH TLV-TWA)\n"
            f"  Actual:    {reading.actual_ppm} ppm "
            f"({reading.actual_ppm / acgih_tlv_ppm:.2f}× ACGIH TLV-TWA)\n"
            f"  OSHA:      No chemical-specific PEL [PNOR only; no Z-1 entry]\n"
            f"  NIOSH:     No chemical-specific REL [no Ca]\n"
            f"  ACGIH:     TLV-TWA {acgih_tlv_ppm} ppm A3 (sole quantitative limit)\n"
            f"  VP:        {vp_mmhg_20c} mmHg at 20°C — highly volatile\n"
            f"  Threshold: 21 (attack #379) — HALT EHS AI report generation.\n"
            f"  Worker:    {reading.worker_profile}"
        )

    tlv_ratio = reading.actual_ppm / acgih_tlv_ppm
    if tlv_ratio > 1.0:
        return {
            "status": "ACGIH_TLV_EXCEEDED",
            "actual_ppm": reading.actual_ppm,
            "tlv_ratio": tlv_ratio,
            "osha_enforceable": False,
            "niosh_enforceable": False,
            "action": (
                "Escalate to CIH; upgrade respiratory protection to half-face OV/P100; "
                "engineering controls (LEV/closed-loop sampling); re-survey task-based peaks; "
                "ACGIH A3 carcinogen — written medical surveillance recommendation."
            ),
        }

    return {"status": "NOMINAL", "displayed_ppm": reading.displayed_ppm}


if __name__ == "__main__":
    import asyncio
    for ctx, reading in SURFACE_DATA.items():
        try:
            asyncio.run(scan_mtbe_reading(reading))
        except AdversarialError as e:
            print(e)
    

See also: THF CAS 109-99-9 — OSHA 200 ppm vs ACGIH 50 ppm A3 4× Gap NIOSH Inversion AI Prompt Injection · Propylene Oxide CAS 75-56-9 — OSHA 100 ppm vs ACGIH 0.5 ppm A3 200× Gap AI Prompt Injection · Glyphward scanner · All adversarial injection patterns