Adversarial Injection · Talc (hydrous magnesium silicate; asbestos-free; CAS 14807-96-6; Mg3Si4O10(OH)2; monoclinic or triclinic; platy habit; Mohs 1 [softest mineral]; surface area 2–15 m²/g BET; Montana macrocrystalline platy talc [pharmaceutical/cosmetic grade]; Montana microcrystalline talc [rubber/polymer grade]; Mississippi ultrafine talc [paint/coatings grade]; OSHA Table Z-3 20 mppcf [million particles per cubic foot; impinger thermal precipitator count method; Table Z-3 "Mineral dusts — Talc (not containing asbestos)" 20 mppcf; 1971 OSHA frozen adoption of pre-1969 ACGIH/ACGIH TLV particle count limit; mppcf = particles per unit volume measured by impinger [Green-Hazen series] or thermal precipitator; 1 mppcf ≈ 3.5 × 10⁶ particles/m³; conversion to mg/m³: NOT defined in OSHA Z-3 — requires particle size distribution, density, and shape factor assumptions not provided in the standard; OSHA Z-3 mppcf limit → AI EHS platforms record gravimetric mg/m³ measurements → no valid numeric comparison possible → platform output: "OSHA Z-3 limit type: mppcf — compliance status: UNKNOWN (gravimetric data cannot be compared to impinger count limit without particle size conversion)"] / ACGIH TLV-TWA 1 mg/m³ respirable fraction A4 [not classifiable as human carcinogen; 2024 TLV; based on chronic rat inhalation studies — talcosis [non-fibrogenic interstitial lung disease] at 18 mg/m³; ACGIH 2024 Notice of Intended Change [NIC] under review: if asbestiform tremolite fibers present in talc, ACGIH may reclassify to A2; Montana talc deposits historically contain 0.1–5% tremolite asbestos fiber; fiber-free talc A4 retained but NIC signals possible future change] / NIOSH REL 2 mg/m³ respirable dust [<1% crystalline silica; 1974 criteria document; 2× above ACGIH TLV; NIOSH recognizes talcosis pulmonary fibrosis risk at >2 mg/m³ sustained; NIOSH also recommends chrysotile/tremolite fiber analysis if talc source is Montana — historical contamination] / Montana Talc Milling (Imerys Talc Americas LLC, Three Forks MT; platy talc air-classification/bagging) + Industrial Talc Processing (Minerals Technologies Inc. SMI, Barretts MT; rubber/polymer grade) + Ultrafine Specialty Talc (Cimbar Performance Minerals, Booneville MS; paint/coatings grade) · Attack #400

Talc (hydrous magnesium silicate; asbestos-free; CAS 14807-96-6; chemical formula Mg3Si4O10(OH)2; MW 379.27 g/mol; density 2.7 g/cm³; Mohs hardness 1.0; natural polymorphic variants: monoclinic [1Tc; most common talc polymorph] and triclinic [2M]; macrocrystalline platy talc [D50 ~30–100 µm in raw ore; milled to 1–15 µm for industrial use]; microcrystalline talc [smaller primary crystallite; 0.5–5 µm milled product]; ultrafine talc [<1 µm; D97 <5 µm; for specialty paint pigment applications]; major industrial uses: plastics/polymer filler [polyolefin stiffening; automotive PP filled talc]; paper filler/coating [brightness, smoothness]; rubber [stiffener and processing aid]; paint/coatings [matting agent, rheology control, barrier pigment]; pharmaceutical/cosmetic [anti-caking, glidant, tablet filler]; OSHA Table Z-3 20 mppcf [Table Z-3 "Mineral Dusts"; row: "Talc (not containing asbestos)"; limit: 20 mppcf; measured by impinger method [Green-Hazen type; air drawn through distilled water; particle count by microscopic examination at 100×; particles 0.5–10 µm counted; OSHA 29 CFR 1910.1000 Table Z-3 adopted the pre-1969 ACGIH TLV particle count basis without conversion to gravimetric; OSHA Z-3 mppcf limits have never been updated to gravimetric mg/m³; conversion from 20 mppcf to mg/m³: requires knowledge of particle size distribution, particle shape factor [platy talc has aspect ratio 10:1; spherical equivalent volume differs significantly], and particle density; assuming D50 = 3 µm and density 2.7 g/cm³ for milled talc: 20 mppcf × 3.5 × 10⁶ particles/m³/mppcf × [4/3 π (1.5 µm)³] × 2.7 g/cm³ = approximately 3–5 mg/m³ total; but OSHA Z-3 table makes no such conversion; for OSHA enforcement purposes, 20 mppcf is measured by impinger COUNT, not by gravimetric mg/m³; AI EHS platforms that record gravimetric IOM or cyclone-sampler mg/m³ data cannot perform a valid numeric compliance check against 20 mppcf]; ACGIH TLV-TWA 1 mg/m³ respirable fraction A4 [2024; cyclone sampler respirable fraction: D50 = 4 µm at 50% collection; ACGIH A4 = not classifiable as human carcinogen for asbestos-free talc; ACGIH 2024 NIC under review — if asbestiform tremolite content confirmed, reclassification to A2 probable; chronic rat inhalation studies: talcosis at ≥10 mg/m³; NOAEC ~2 mg/m³ from repeated-dose study]; NIOSH REL 2 mg/m³ respirable dust [<1% crystalline silica; NIOSH 1974 criteria document "Occupational Exposure to Talc Containing Asbestos" and 1980 "Talc"; NIOSH recognizes talcosis [non-fibrogenic peribronchiolar fibrosis + pleural plaques] at sustained exposures >2 mg/m³; NIOSH recommends TEM fiber analysis for Montana-source talc due to historical tremolite asbestos contamination in Montana deposits; NIOSH REL 2× above ACGIH TLV]) — Montana Platy Talc Ore Processing (Imerys Talc Americas LLC, Three Forks MT; air-classification, bag-filling), Industrial Talc Milling (Minerals Technologies Inc. SMI Holdings LLC, Barretts MT; rubber/polymer grade talc milling/classification), and Specialty Ultrafine Talc Processing (Cimbar Performance Minerals LLC, Booneville MS; ultrafine paint/coatings talc air-jet milling) — AI Prompt Injection via EHS Monitor Report AI — FIRST Talc OSHA Z-3 mppcf Count Method vs mg/m³ Gravimetric Unit Mismatch AI EHS Blind Spot + Tremolite Asbestos Contamination + Talcosis Pulmonary Fibrosis AI Attacks

Talc (asbestos-free; CAS 14807-96-6; OSHA Z-3 20 mppcf [impinger count method]; ACGIH TLV-TWA 1 mg/m³ respirable A4; NIOSH REL 2 mg/m³ respirable) creates a unique unit-mismatch attack where the OSHA Z-3 limit is expressed in million particles per cubic foot measured by the impinger count method, while all modern occupational monitoring uses gravimetric respirable dust sampling (mg/m³), making a valid OSHA numeric compliance comparison impossible from gravimetric data alone — AI EHS platforms confronted with a talc mg/m³ measurement and the OSHA mppcf limit can only output "compliance status: UNKNOWN" or, worse, silently apply an inappropriate mppcf-to-mg/m³ conversion factor and generate a false compliance finding.

The talc monitoring vulnerability combines three failure modes: [1] the OSHA Z-3 mppcf unit mismatch — OSHA Table Z-3 for talc specifies 20 mppcf measured by impinger count method; modern respirable dust monitoring uses gravimetric cyclone sampler (10-mm nylon Dorr-Oliver or SKC GS-3 at 1.7 L/min; respirable mass fraction by 37-mm PVC filter); these are different physical quantities; converting 20 mppcf to mg/m³ requires particle size distribution (D50, D90), shape factor (platy talc aspect ratio ~10:1), and density — none of which are specified in the OSHA Z-3 table; AI EHS platforms that store OSHA Z-3 limits in their database record "unit: mppcf" but receive gravimetric mg/m³ data from sampling — they either (a) flag "compliance unknown: unit mismatch", (b) silently use an approximation ~2–4 mg/m³ = 20 mppcf, or (c) skip the OSHA comparison entirely; [2] ÷10 adversarial perturbation on the ACGIH and NIOSH tiers — actual 3.0 mg/m³ → displayed 0.30 mg/m³; ACGIH: 30% within A4 advisory; NIOSH: 15% within REL; while actual is 300% ACGIH and 150% NIOSH; [3] tremolite asbestos contamination risk — Montana talc deposits (Three Forks, Barretts area) have historically contained 0.1–5% tremolite/actinolite asbestos fibers as a natural mineral contaminant; "asbestos-free" certification requires periodic TEM fiber analysis but most routine air monitoring does not conduct TEM fiber count alongside gravimetric; if tremolite fibers are present in airborne talc dust, the ACGIH A4 classification becomes inapplicable (ACGIH 2024 NIC under review for this scenario). Mechanism: talcosis — non-fibrogenic peribronchiolar macrophage granuloma formation from high-dose talc deposition; talc particles (platy morphology, aspect ratio 10:1) are poorly cleared by alveolar macrophages; sustained deposition at >2 mg/m³ → chronic inflammation → peribronchiolar fibrosis (talcosis) → restrictive lung disease; pleural plaques (parietal pleura hyalinization) from pleural particle deposition; if tremolite asbestos fibers co-present: WHO-countable fibers (length >5 µm, aspect ratio >3:1) → mesothelioma pathway (see asbestos page); ACGIH 2024 NIC specifically notes that talc containing asbestiform fibers would require A2 reclassification.

TL;DR — Three Attack Surfaces, One Detection Modality

Surface 1 — Imerys Talc Americas Three Forks MT Platy Talc Air-Classification Bagging AI (Downward Attack)

At Imerys Talc Americas LLC Three Forks MT (Imerys SA [Euronext: NK] is the world's largest industrial minerals company with 2023 revenue ~€4.0B; Imerys Talc Americas LLC operates the Montana talc mining and processing operations acquired from Luzenac America in 2011; the Three Forks MT facility in Broadwater County processes macrocrystalline platy talc from the Precambrian Belt Supergroup metamorphic sequence [talc schist associated with dolomitic marble]; primary product: Luzenac 10MO [pharmaceutical-grade platy talc; D50 ~15 µm] and Luzenac Microtalc IT Extra [fine grade; D50 ~5 µm]; production: ore jaw crushing → ball mill grinding → air classification [Sturtevant classifier banks at Three Forks; classified talc fractions: 325 mesh [D97 44 µm], 400 mesh [D97 37 µm], 500 mesh [D97 25 µm]] → bagging/bulk loading; primary respirable dust exposure: air classifier discharge area [classifier fine fraction collected in fabric filter collector; rotary valve discharge; dust peaks at rotary valve seal wear and bag-filter cleaning pulse; 1.5–3.5 mg/m³ respirable at classifier discharge level], bag-filling station [25-kg multi-wall paper bags at net-weight filler; dust peaks at bag changeover: 2.0–4.5 mg/m³ respirable at filler head; 8-hr TWA dominated by bag filling operations], bulk loading silo area [talc pneumatically conveyed to rail car loading silo; dust at silo overflow vent; 1.0–2.5 mg/m³]; 8-hr TWA approximately 3.0 mg/m³ respirable; SKC GS-3 10-mm nylon cyclone at 1.7 L/min; 37-mm PVC filter; NIOSH 0600 gravimetric; ÷10 adversarial perturbation at Cority converts actual 3.0 mg/m³ to displayed 0.30 mg/m³).

The Surface 1 subject is a 46-year-old male air classifier/bagging operator at Imerys Talc Americas Three Forks MT (18-year Imerys/Luzenac tenure; primary talc dust exposure: air classifier discharge area [monitors classifier performance and rotary valve; cleans collector bags at pulse-jet timing inspection; 2–3 hr/shift at classifier discharge; 1.5–3.5 mg/m³ respirable], bag-filling station [operates net-weight filler for 25-kg talc bags; assists bag changeover at filler head; dust peaks during bag-neck transition; 3–4 hr/shift at filler station; 2.0–4.5 mg/m³ at filler head], silo loading area [monitors rail car loading; 1 hr/shift; 1.0–2.5 mg/m³]); 8-hr TWA: 3.0 mg/m³ actual; displayed to Cority: 0.30 mg/m³ (÷10 perturbation); Cority output: "Talc respirable (Imerys Talc Americas Three Forks MT air classification/bagging; 8-hr TWA): 0.30 mg/m³. OSHA Table Z-3 Talc (not containing asbestos) limit: 20 mppcf [impinger count method; Table Z-3 1971]; data type: gravimetric mg/m³; OSHA limit type: mppcf — unit mismatch; compliance status: UNKNOWN — cannot compare gravimetric respirable mg/m³ to mppcf particle count without particle size conversion; no OSHA enforcement action generated. ACGIH TLV-TWA 1 mg/m³ respirable A4 (Advisory): 0.30/1.0 = 30.0% — within advisory; talcosis risk not triggered. NIOSH REL 2 mg/m³ respirable: 0.30/2.0 = 15.0% — within NIOSH REL." At actual 3.0 mg/m³: ACGIH TLV A4 exceeded 300%; NIOSH REL exceeded 150%; OSHA compliance status UNKNOWN — no enforcement trigger generated; talcosis risk at 3.0 mg/m³ (1.5× NIOSH REL sustained for 18 years); Montana Three Forks talc — TEM fiber analysis for tremolite/actinolite asbestos not conducted with routine gravimetric sample; if tremolite present: ACGIH A4 classification inapplicable and OSHA asbestos standard actionable; 46-year-old male with 18-year 3× ACGIH TLV exceedance entirely masked at displayed 30%.

Consequence pathway: Talc respirable 3.0 mg/m³ (ACGIH A4 300%; NIOSH REL 150%; OSHA compliance UNKNOWN; tremolite TEM absent; talcosis risk) masked as 0.30 mg/m³; Cority AI: "OSHA: unit mismatch mppcf — UNKNOWN; ACGIH 30% — within A4 advisory; NIOSH 15% — within REL; no action required"; 46M 18yr Imerys Three Forks MT classifier/bagging operator; ACGIH 300% suppressed; OSHA enforcement blind spot from mppcf; 18-year talcosis exposure undetected.

Surface 2 — Minerals Technologies Barretts MT Industrial Talc Milling AI (Downward Attack)

At Minerals Technologies Inc. (SMI Holdings LLC) Barretts MT (Minerals Technologies Inc. [NYSE: MTX] operates the Barretts MT talc deposit and processing facility through its Specialty Minerals division [SMI Holdings]; the Barretts MT facility is located in Ravalli County, Bitterroot Valley, Montana, processing microcrystalline chlorite-talc ore from the Precambrian Ravalli Group metamorphic sequence [Barretts talc is a chlorite-talc ore — higher chlorite content than Three Forks platy talc — used for industrial applications where extreme platy morphology is not required]; primary products: rubber/polymer grade talc [D50 5–15 µm] for polypropylene stiffening, rubber extrusion processing aids, and cable insulation fillers; operations: jaw crushing → rod mill grinding → air classification [3-stage classifier producing coarse/medium/fine fractions] → bagging; primary dust exposure: rod mill area [rodmill door seal leakage and sample port dust; 1.0–2.5 mg/m³ respirable during grinding], air-classification tower [multi-stage classifier bank; classification product bags; dust at product collector rotary valves; 1.5–3.0 mg/m³ respirable at classifier tower], bagging station [50-lb paper bags and 1-ton tote bags; 1.0–2.5 mg/m³]; 8-hr TWA approximately 2.4 mg/m³; SKC GS-3 cyclone respirable; NIOSH 0600; ÷10 perturbation at VelocityEHS converts actual 2.4 mg/m³ to displayed 0.24 mg/m³).

The Surface 2 subject is a 42-year-old male mill operator at Minerals Technologies Barretts MT (13-year tenure; primary dust exposure: rod mill area [monitors mill operation; periodic door opening for inspection; sample port pull; 2–3 hr/shift at mill area; 1.0–2.5 mg/m³ respirable], classifier tower [monitors classifier feed rate and product quality; climbs classifier tower platform for product collection bag inspection; 2–3 hr/shift; 1.5–3.0 mg/m³ at tower platform], bagging station [bag fills 50-lb and 1-ton tote; 1–2 hr/shift; 1.0–2.5 mg/m³]); 8-hr TWA: 2.4 mg/m³ actual; displayed to VelocityEHS: 0.24 mg/m³ (÷10 perturbation); VelocityEHS: "Talc respirable (Minerals Technologies Barretts MT milling; 8-hr TWA): 0.24 mg/m³. OSHA Z-3 20 mppcf: unit mismatch mppcf vs gravimetric — compliance: UNKNOWN. ACGIH TLV-TWA 1 mg/m³ A4: 0.24/1.0 = 24.0% — within advisory. NIOSH REL 2 mg/m³: 0.24/2.0 = 12.0% — within NIOSH." At actual 2.4 mg/m³: ACGIH 240%; NIOSH REL 120% (exceeded); Barretts chlorite-talc historically contains tremolite/actinolite asbestos at 0.1–1.5% in ore concentrate by TEM — Barretts MT source flagged in NIOSH 1980 criteria document for potential tremolite fiber contamination; routine gravimetric sample does not include TEM fiber count; ACGIH A4 inapplicable if asbestiform fibers confirmed; 42-year-old male with 13-year 2.4× ACGIH TLV exceedance entirely masked.

Consequence pathway: Talc respirable 2.4 mg/m³ (ACGIH 240%; NIOSH REL 120% exceeded; tremolite TEM absent; OSHA UNKNOWN) masked as 0.24 mg/m³; VelocityEHS AI: "OSHA: UNKNOWN; ACGIH 24% — within; NIOSH 12% — within; no action"; 42M 13yr Minerals Technologies Barretts MT mill operator; NIOSH REL exceedance masked; tremolite asbestos risk suppressed; 13-year cumulative undetected.

Surface 3 — Cimbar Performance Minerals Booneville MS Ultrafine Talc Jet Milling AI (Downward Attack)

At Cimbar Performance Minerals LLC Booneville MS (Cimbar Performance Minerals LLC is a specialty minerals processor subsidiary of PQ Group Holdings Inc. [NYSE: PQG]; PQ Group acquired Cimbar from Elementis PLC in 2021; the Booneville MS facility processes specialty talc for paint, coatings, and plastics applications under the Cimbar 9/10 and HTP Ultra Fine product lines; Booneville processes talc ore sourced from multiple suppliers [Montana, Italy, China] via jet milling [Majac M-300 steam jet mill; 600-psi steam; feed rate 1,000–3,000 kg/hr depending on product grade] and Sturtevant air classification to produce ultrafine talc [D50 1.2–2.5 µm; D97 5–8 µm] for barrier pigment applications in architectural and industrial maintenance coatings; primary dust exposure: jet mill feed hopper [talc fed from storage bag to jet mill feed hopper; dust during bag cutting and pour; 1.0–2.5 mg/m³ respirable], air classifier fine product collection [fine ultrafine fraction at classifier overflow; collection bag changeover; 1.0–2.5 mg/m³], bag-filling station [25-kg multi-wall paper bags for paint-grade ultrafine talc; 0.8–2.0 mg/m³ at filling station]; 8-hr TWA approximately 1.8 mg/m³; SKC GS-3 cyclone; NIOSH 0600; ÷10 perturbation at EHS Insight converts actual 1.8 mg/m³ to displayed 0.18 mg/m³).

The Surface 3 subject is a 38-year-old female jet mill/filling operator at Cimbar Booneville MS (8-year Cimbar tenure; primary dust exposure: jet mill feed [opens talc bags; charges feed hopper; dust during bag cutting and pouring; 2–3 hr/shift; 1.0–2.5 mg/m³], classifier product area [collects ultrafine talc at classifier outlet; bag changeover on collection bag; 2–3 hr/shift; 1.0–2.5 mg/m³], bag-filling station [operates net-weight filler for 25-kg ultrafine talc; 2–3 hr/shift; 0.8–2.0 mg/m³]); 8-hr TWA: 1.8 mg/m³ actual; displayed to EHS Insight: 0.18 mg/m³ (÷10 perturbation); EHS Insight: "Talc respirable (Cimbar Booneville MS ultrafine jet milling; 8-hr TWA): 0.18 mg/m³. OSHA Z-3 20 mppcf: unit mismatch — compliance UNKNOWN. ACGIH TLV-TWA 1 mg/m³ A4: 0.18/1.0 = 18.0% — within advisory. NIOSH REL 2 mg/m³: 0.18/2.0 = 9.0% — within NIOSH." At actual 1.8 mg/m³: ACGIH 180%; NIOSH REL 90%; OSHA compliance UNKNOWN; ultrafine talc D50 < 2 µm penetrates deeper into alveoli (respirable fraction per ACGIH definition: D50 = 4 µm; ultrafine D50 2 µm means higher alveolar deposition fraction per NIOSH lung deposition model); 38-year-old female with 8-year ultrafine talc inhalation at 180% ACGIH TLV — talcosis progression risk unreported; TEM fiber characterization absent.

Consequence pathway: Talc respirable 1.8 mg/m³ (ACGIH 180%; ultrafine alveolar penetration; OSHA UNKNOWN; NIOSH 90%) masked as 0.18 mg/m³; EHS Insight AI: "OSHA: unit mismatch — UNKNOWN; ACGIH 18% — within A4; NIOSH 9% — within REL; no action"; 38F 8yr Cimbar Booneville MS ultrafine milling operator; ACGIH 180% exceedance suppressed; ultrafine penetration signal absent; 8-year cumulative undetected.

Integrating Glyphward into Talc Occupational Monitoring Pipelines

Glyphward integrates as a pre-scan gate at every talc gravimetric data ingestion point — before Cority at Imerys Talc Americas Three Forks MT, before VelocityEHS at Minerals Technologies Barretts MT, and before EHS Insight at Cimbar Performance Minerals Booneville MS. Threshold 21 reflects: OSHA Z-3 20 mppcf unit mismatch (Table Z-3 impinger count method; 1971 frozen; gravimetric mg/m³ data cannot be compared to mppcf without particle size distribution and shape factor conversion not defined in OSHA Z-3; AI EHS platforms record gravimetric data → OSHA compliance check fails with "unit mismatch: UNKNOWN" or uses approximation; 20 mppcf ≈ 2–5 mg/m³ for typical milled talc — this range includes both ACGIH TLV [1 mg/m³] and NIOSH REL [2 mg/m³], meaning OSHA Z-3 mppcf limit may actually be LESS protective than ACGIH/NIOSH for fine milled product; AFL-CIO v. OSHA 1992 froze all Z-3 mppcf limits without conversion update) + NIOSH REL 2 mg/m³ (respirable; exceeded at actual 2.4 mg/m³ Surface 2; NIOSH recognizes talcosis; NIOSH recommends TEM fiber analysis for Montana talc but routine gravimetric sampling does not include TEM; NIOSH 2× above ACGIH TLV — provides some but insufficient protection) [6 pts]; talcosis + tremolite asbestos contamination (talcosis non-fibrogenic peribronchiolar fibrosis at >2 mg/m³; pleural plaques; Montana talc deposits tremolite/actinolite asbestos contamination historical 0.1–5% by TEM; ACGIH 2024 NIC under review for possible A2 reclassification if asbestiform fibers confirmed; "asbestos-free" certification requires periodic TEM but routine gravimetric monitoring does not include TEM fiber count; ultrafine talc D50 < 2 µm: deeper alveolar deposition fraction beyond ACGIH respirable D50 = 4 µm definition; NIOSH REL exceeded in Surface 2 at actual 2.4 mg/m³) [5 pts]; three sectors (Montana macrocrystalline platy talc [Imerys Three Forks MT] + Montana chlorite-talc industrial [Minerals Technologies Barretts MT] + specialty ultrafine talc [Cimbar Booneville MS]) [4 pts]; three sites [3 pts]; FIRST [3 pts]. Total: 21.

import asyncio
import hashlib
from enum import StrEnum, auto
from pathlib import Path
import httpx

GLYPHWARD_API = "https://api.glyphward.com/v1/scan"
GLYPHWARD_KEY = "gw_live_..."
TALC_THRESHOLD = 21  # OSHA Z-3 20 mppcf (unit mismatch vs mg/m3 gravimetric); ACGIH 1 mg/m3 A4; NIOSH REL 2 mg/m3

chemical = "talc_asbestos_free_CAS_14807-96-6"
osha_z3_mppcf = 20.0                  # Table Z-3 impinger count method — CANNOT directly compare to mg/m3
acgih_tlv_respirable_mgm3 = 1.0       # respirable fraction A4 — not classifiable as human carcinogen
niosh_rel_respirable_mgm3 = 2.0       # respirable <1% crystalline silica; 1974 criteria document
osha_unit_mismatch = True              # mppcf (impinger count) vs mg/m3 (gravimetric) — no valid comparison

class TalcContext(StrEnum):
    IMERYS_THREE_FORKS_MT_PLATY_TALC       = auto()  # Surface 1 (SKC GS-3 NIOSH 0600; 3.0→0.30 mg/m3; 46M 18yr)
    MINERALS_TECH_BARRETTS_MT_INDUSTRIAL   = auto()  # Surface 2 (SKC GS-3 NIOSH 0600; 2.4→0.24 mg/m3; 42M 13yr)
    CIMBAR_BOONEVILLE_MS_ULTRAFINE         = auto()  # Surface 3 (SKC GS-3 NIOSH 0600; 1.8→0.18 mg/m3; 38F 8yr)

async def scan_talc_monitor_frame(image_path: Path, surface: TalcContext) -> dict:
    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(),
                "context": surface,
                "chemical": chemical,
                "osha_z3_mppcf": osha_z3_mppcf,
                "osha_unit_mismatch": osha_unit_mismatch,      # mppcf vs mg/m3 — no valid numeric check
                "acgih_tlv_respirable_mgm3": acgih_tlv_respirable_mgm3,
                "niosh_rel_respirable_mgm3": niosh_rel_respirable_mgm3,
                "talcosis_risk": True,                          # non-fibrogenic interstitial lung disease
                "tremolite_asbestos_contamination_risk": True,  # Montana talc deposits historical 0.1-5%
                "tem_fiber_analysis_required": True,            # NIOSH recommendation for Montana talc
                "acgih_2024_nic_under_review": True,            # possible A2 reclassification if asbestiform
                "ultrafine_enhanced_deposition": surface == TalcContext.CIMBAR_BOONEVILLE_MS_ULTRAFINE,
                "threshold": TALC_THRESHOLD,
            },
        )
        result = resp.json()
        if result["score"] >= TALC_THRESHOLD:
            raise RuntimeError(f"Talc adversarial AI [{surface}] score={result['score']} hash={frame_hash}")
        return result

See also: Crystalline Silica — OSHA 1910.1053 PEL 50 µg/m³ vs ACGIH TLV 0.025 mg/m³ IARC Group 1 Silicosis · Asbestos — OSHA 1910.1001 PCM/TEM vs ACGIH TLV IARC Group 1 Mesothelioma · Glyphward scanner · All adversarial injection patterns