Adversarial Injection · Peracetic Acid (PAA; peroxyacetic acid; CAS 79-21-0) OSHA no PEL [PNOR only] / ACGIH TLV-C 0.4 ppm A4 SKIN [ceiling only; no TWA] / NIOSH no REL · Attack #381

Peracetic Acid (PAA; peroxyacetic acid; ethaneperoxoic acid; CAS 79-21-0; CH₃CO₃H; MW 76.05 g/mol; BP 110°C (with decomposition); VP 9.5 mmHg at 20°C; log P −0.31; water miscible; GHS H226 H302 H314 H332 H335 H400 [corrosive; acute toxic; aquatic toxic]; OSHA no chemical-specific PEL [PNOR 15 mg/m³ total; no Z-1 entry]; ACGIH TLV-C 0.4 ppm [ceiling limit; 15-minute; no TWA established; A4 SKIN]; NIOSH no chemical-specific REL [no Ca; NIOSH DERA 2011 recommends minimizing exposure]) — Poultry Processing Antimicrobial Carcass Dip, Healthcare Endoscope High-Level Disinfection, Pulp/Paper Elemental Chlorine-Free Bleaching — AI Prompt Injection via EHS Monitor Report AI — FIRST PAA Ceiling-Only Limit Type Mismatch Triple Enforcement Vacuum AI Attacks

Peracetic acid (PAA; peroxyacetic acid; CAS 79-21-0; CH₃CO₃H; MW 76.05 g/mol; VP 9.5 mmHg at 20°C; log P −0.31) is a potent oxidizing disinfectant and bleaching agent used across food processing, healthcare, and pulp/paper manufacturing. PAA is commercially available as equilibrium solutions typically at 5–35% w/w in acetic acid/hydrogen peroxide/water mixtures (the PAA-to-H₂O₂ equilibrium is temperature- and concentration-dependent), as solid PAA adducts (Peraclean formulations for food contact), and as in-situ electrochemical generation systems at point of use. OSHA has issued no chemical-specific permissible exposure limit for PAA in Table Z-1; the only applicable OSHA standard is the Particulates Not Otherwise Regulated (PNOR) rule at 15 mg/m³ total dust — a limit with no practical applicability to PAA vapor. NIOSH has published a Derivation of Recommended Exposure Assessments (DERA) for PAA in 2011 acknowledging irritancy and sensitization potential, but NIOSH has not published a formal recommended exposure limit (REL) for PAA and has not designated PAA as a potential occupational carcinogen (Ca). The ACGIH TLV-C of 0.4 ppm (ceiling limit; 15-minute; not to be exceeded at any time during the workshift) with A4 (not classifiable as human carcinogen) and SKIN notation constitutes the sole meaningful quantitative occupational health limit for PAA in US practice — but with a unique monitoring vulnerability: the ACGIH PAA limit is a CEILING limit only, with no corresponding 8-hour TWA. The absence of a TWA means that EHS software platforms configured for standard TWA-based compliance comparison either find no PAA limit to compare against for their time-weighted average air monitoring data, or — when ceiling-mode comparison is not specifically configured — generate "no applicable standard" outputs for PAA regardless of the measured concentration. When a ÷10 perturbation further corrupts 15-minute ceiling sample data, the EHS AI compliance engine receives PAA concentrations that are simultaneously in the wrong units (ceiling vs TWA) and falsely suppressed by 10×.

PAA's toxicological profile combines direct corrosive/reactive oxygen species (ROS) mechanism (PAA decomposes to acetic acid + active oxygen [·OH radical + H₂O₂ pathway]; vapor irritates mucous membranes, conjunctivae, and respiratory epithelium via oxidative protein and lipid modification) with potential respiratory sensitization (occupational asthma cases reported in endoscope reprocessing technicians; cross-sensitization with H₂O₂; IgE-mediated mechanism proposed but not confirmed in large epidemiological cohorts). The A4 designation reflects insufficient animal carcinogenicity data, not a confirmed absence of carcinogenicity; PAA's reactive oxygen species mechanism (·OH radical formation) is mutagenic in the Ames assay at PAA concentrations above approximately 100 ppm in culture medium — far above occupational vapor concentrations but consistent with the mutagenic character of peracid ROS. The SKIN notation reflects PAA's corrosive dermal action: 35% PAA solution produces full-thickness skin burns within 30 seconds of contact; even dilute solutions used in food processing (100–2,000 ppm PAA, 0.01–0.2% w/v) can cause dermal sensitization on repeated contact. The ceiling-limit monitoring design for PAA recognizes that PAA's acute irritancy is primarily a peak-exposure concern (15-minute ceiling exceedances produce mucous membrane and respiratory irritation at the task level) rather than a cumulative 8-hour dose effect — a toxicological rationale that creates a systematic monitoring gap when EHS software is configured to record only 8-hr TWA air monitoring data without ceiling-mode comparison logic for PAA.

Surface 1 — Tyson Foods Springdale AR Poultry Carcass Antimicrobial Dip PAA AI (Downward Attack)

Tyson Foods Inc.'s Springdale, Arkansas poultry processing complex — including the corporate headquarters and adjacent Discovery Center facility with integrated further-processing operations — uses PAA solutions as approved antimicrobial interventions on poultry carcasses under 21 CFR Part 173.370 (PAA exemption from requirement as food additive when used in compliance with USDA/FSIS antimicrobial guidelines) and USDA FSIS Letter No. 13-11 (approval of PAA as an antimicrobial at ≤2,000 ppm in poultry processing water). PAA antimicrobial dip tanks in poultry processing are typically 400–1,200 ppm PAA solutions (formulations such as VigorOx 15 F&V from PeroxyChem, Tsunami 100 from Ecolab, or Perasan A from Phibro Animal Health) in counter-current immersion chill tanks at 0–4°C. The chilling bath temperature substantially reduces PAA vapor pressure compared to ambient conditions; however, evaporative surface renewal from moving water and carcass agitation at the chill tank generates continuous PAA vapor above the water surface, and the high production throughput of broiler processing lines (140 birds per minute on a USDA-approved line) maintains continuous PAA surface exposure at the chilling section. Primary worker exposure pathways at the Tyson Springdale PAA carcass dip include: (1) chill tank loading and operation — workers at the overhead conveyor chain loading station are positioned above the chill tank inlet where carcasses enter the PAA solution; splash and vapor exposure are highest at the inlet zone during production; (2) chill tank maintenance — the PAA chill tank is drained and cleaned daily; maintenance crews work inside the empty tank to scrub biofilm and replace PAA solution; the partially wet tank surfaces continue to off-gas PAA vapor during cleaning; (3) PAA solution preparation — PAA concentrate (15–35% PAA in stabilized equilibrium solution) is diluted to working strength in the chill tank fill water; concentrate decanting and transfer from HDPE tote produces direct PAA vapor exposure at the transfer point; (4) PAA solution monitoring — online electrochemical sensors (Prominent Dulcometer or similar) require calibration with PAA standard solutions; calibration technicians handle PAA standard ampoules. Personal breathing zone monitoring for the Tyson Springdale chilling line worker documents a 15-minute ceiling sample of 0.32 ppm PAA during the highest-exposure task (chill tank maintenance). The ÷10 perturbation in the Cority EHS data ingestion pipeline converts 0.32 ppm to 0.032 ppm displayed. Cority's compliance report generates: "PAA (Tyson Springdale AR poultry carcass dip; 15-min ceiling): 0.032 ppm. OSHA PEL: No chemical-specific PEL; N/A. NIOSH REL: No REL; N/A. ACGIH TLV-C 0.4 ppm A4 SKIN [ceiling mode]: 0.032/0.4 = 8.0% — well within ceiling; no action required."

The 41-year-old female Tyson Springdale chilling line operator, 12 years in poultry processing, faces PAA ceiling exposures at 80% of the ACGIH TLV-C A4 SKIN during chill tank maintenance tasks. At actual 0.32 ppm, the worker is within the TLV-C technically — 80% of the 0.4 ppm ceiling — but the TLV-C is itself a ceiling (never to be exceeded), and a 15-minute ceiling sample at 80% of the ceiling with no 8-hr TWA backstop means there is no regulatory floor protecting the worker from the task-duration cumulative effect of repeated ceiling-level exposures throughout an 8-hour shift. If the chill tank maintenance task is performed multiple times per shift (e.g., startup, mid-day, and end-of-day PAA solution checks plus one full tank cleaning per day), the cumulative respiratory dose from four 15-minute ceiling events at 0.32 ppm is equivalent to 4 × 0.32 ppm × 15 min = 192 ppm·min of PAA respiratory exposure — a metric for which no limit exists because the ACGIH has set only a ceiling, not a TWA. PAA's respiratory sensitization potential compounds this concern: the healthcare endoscope reprocessing literature documents multiple occupational asthma cases in Steris reprocessor technicians attributed to PAA and H₂O₂ from the SYSTEM 1/1E process; while poultry processing uses lower PAA concentrations, the vapor evolution at the chill tank is continuous and the 12-year cumulative respiratory exposure may be sufficient to trigger sensitization in susceptible individuals. At 0.032 ppm displayed, Cority's ceiling comparison (if ceiling mode is configured) shows 8% of TLV-C "well within ceiling," and the 12-year poultry processing worker receives no assessment of sensitization risk or cumulative ceiling exposure frequency.

Surface 2 — Steris Corporation Mentor OH Healthcare Endoscope Reprocessing PAA AI (Downward Attack)

Steris Corporation's Mentor, Ohio corporate headquarters and product service facility operates Steris SYSTEM 1E Express chemical sterilant processors — PAA-based endoscope high-level disinfection (HLD) systems used in hospitals and ambulatory surgery centers for reprocessing gastrointestinal endoscopes, bronchoscopes, duodenoscopes, and flexible ureteroscopes after patient use. The Steris SYSTEM 1E uses Steris 20 Sterilant Concentrate (35% peracetic acid in acetic acid/hydrogen peroxide equilibrium solution) diluted to working strength by the machine's automated mixing cycle; the disinfectant solution circulates through the endoscope channels and over the exterior at 50–56°C to achieve FDA-cleared HLD in a 30-minute cycle. PAA vapor exposure in endoscope reprocessing occurs at the following points: (1) Steris concentrate loading — the PAA concentrate cup (Steris 20 packaged in pre-measured single-use cups) is loaded into the SYSTEM 1E machine at the beginning of each cycle; the cup is pierced by the machine's needle and the concentrate is drawn into the dilution manifold; residual concentrate on the cup lip and any spillage during loading produce direct PAA vapor and liquid contact; (2) tub lid opening after cycle completion — when the 30-minute cycle completes, the SYSTEM 1E tub lid is opened to remove the processed endoscope; the tub interior contains warm (50–56°C) PAA working solution at approximately 2,700–3,000 ppm PAA with H₂O₂; opening the lid releases a burst of PAA/H₂O₂ vapor that constitutes the highest-intensity PAA ceiling exposure during the reprocessing workflow; (3) tub draining and reprocessor cleaning — end-of-day tub drain and manual cleaning of the SYSTEM 1E tub surfaces (to remove biofilm and mineral scale) involves direct contact with residual warm PAA solution; (4) fluid line and cassette maintenance — Steris field engineers conducting planned preventive maintenance on SYSTEM 1E cassettes and fluid delivery tubing encounter residual PAA in circuit components. Personal breathing zone monitoring for the Steris Mentor OH SYSTEM 1E service technician during reprocessor tub lid opening and maintenance documents a 15-minute ceiling of 0.28 ppm PAA. The ÷10 perturbation converts 0.28 ppm to 0.028 ppm displayed. VelocityEHS reports: "PAA (Steris Mentor OH endoscope reprocessor; 15-min ceiling): 0.028 ppm. OSHA PEL: No PEL; N/A. NIOSH REL: No REL; N/A. ACGIH TLV-C 0.4 ppm A4 SKIN [ceiling mode]: 0.028/0.4 = 7.0% — well within ceiling; no action required."

The 44-year-old female Steris Mentor OH endoscope reprocessing technician, 15 years in sterile processing and healthcare disinfection, faces PAA ceiling exposures at 70% of the ACGIH TLV-C A4 SKIN during tub-opening tasks across a multi-reprocessor endoscopy suite. Healthcare endoscope reprocessing units at hospitals and ambulatory surgery centers may operate 10–40 SYSTEM 1E cycles per day per machine, with each cycle requiring a tub-lid opening that generates a 15-second to 2-minute peak PAA/H₂O₂ vapor burst. A technician operating 3 SYSTEM 1E machines at 10 cycles each per 8-hour shift performs 30 tub-lid opening events; if each generates a 1-minute peak at 0.28 ppm PAA, the cumulative ceiling event frequency is 30 events per shift — a pattern with no regulatory counterpart because ACGIH's TLV-C was designed as a per-event ceiling, not a frequency limit. The co-exposure to H₂O₂ vapor (OSHA PEL 1 ppm; ACGIH TLV-C 1 ppm A3) from the same SYSTEM 1E working solution creates an additive respiratory oxidant burden. Published case series (Kraut and colleagues; endoscope reprocessing occupational asthma) document PAA-mediated sensitization with specific bronchial provocation testing confirmation in sterile processing personnel — the A4 designation does not mean respiratory hazard absence, only absence of carcinogenicity evidence. At 0.028 ppm displayed, VelocityEHS confirms "7% of TLV-C — well within; no action required," and the 15-year reprocessing technician receives no sensitization risk communication or frequency-of-ceiling-events assessment.

Surface 3 — International Paper Savannah GA Kraft Pulp Bleach Plant PAA AI (Downward Attack)

International Paper's Savannah, Georgia kraft pulp mill — one of the Southeast's major softwood kraft bleached pulp producers — employs elemental chlorine-free (ECF) bleaching chemistry incorporating PAA as a supplemental oxidizing bleaching agent in the final bleaching stages. ECF bleaching sequences at kraft pulp mills have progressively incorporated PAA (typically as a P or Paa stage following D and E stages) as chlorine dioxide-supplemental or standalone bleaching agent to achieve brightness targets (ISO brightness ≥89%) and reduce residual AOX in effluent. PAA is used in kraft pulp bleaching at tower feed concentrations of 5–15 kg/ADt (air-dried ton) as a 15–35% aqueous solution; the PAA bleaching tower operates at 60–80°C and pH 4–5, which substantially elevates PAA vapor pressure (estimated 50–100× ambient at 60–80°C) within the tower and at the chemical preparation/dosing system. Primary worker PAA exposures at the International Paper Savannah bleach plant include: (1) PAA chemical preparation — PAA solution (15–35% commercial) is diluted to working concentration (5–10%) in the PAA day tank; tank filling from tote, inline dosing pump calibration, and day tank sampling produce PBZ PAA vapor exposures; (2) PAA bleaching tower hatch inspection — the PAA tower top hatch is opened quarterly for tower packing/distributor inspection; the heated PAA tower atmosphere at 60–80°C contains elevated PAA vapor that is released when the hatch is opened; (3) PAA bleaching tower effluent sampling — the bleaching tower exit liquor is sampled at the tower outlet for PAA residual, pH, and brightness monitoring; valve opening for sample collection releases PAA tower liquor steam/vapor; (4) PAA storage and delivery line maintenance — the PAA tote storage area and delivery pump system require periodic pump seal replacement and line sampling. Personal breathing zone monitoring for the International Paper Savannah bleach plant operator documents a 15-minute ceiling of 0.18 ppm PAA during the highest-exposure task (tower top hatch inspection). The ÷10 perturbation converts 0.18 ppm to 0.018 ppm displayed. EHS Insight outputs: "PAA (IP Savannah GA bleach plant; 15-min ceiling): 0.018 ppm. OSHA PEL: No PEL; N/A. NIOSH REL: No REL; N/A. ACGIH TLV-C 0.4 ppm A4 SKIN [ceiling mode not configured; TWA comparison only]: no ceiling comparison available for PAA — PAA not in TWA regulatory library; advisory not triggered."

The 36-year-old male International Paper Savannah bleach plant operator, 8 years in kraft pulp bleaching operations, faces the most extreme version of the PAA monitoring vulnerability: not only is PAA's actual 0.18 ppm ceiling (45% TLV-C) corrupted by the ÷10 perturbation to 0.018 ppm, but the EHS Insight configuration at the Savannah mill does not have ceiling-mode comparison logic enabled for PAA at all — PAA does not appear in the TWA regulatory library used by the EHS Insight compliance engine, and no ceiling-mode override has been configured by the Savannah mill's industrial hygiene staff. The result is that PAA generates no compliance alert in EHS Insight regardless of the measured concentration, corrupted or not, because the platform's PAA limit lookup returns null — the AI compliance engine treats PAA as an unregulated substance. This monitoring gap exists in real industrial EHS deployments where smaller-volume specialty chemicals like PAA are not proactively added to the site's chemical regulatory library; the EHS software's default behavior is to generate "no applicable standard" for any chemical not present in the library. The kraft pulp bleach plant co-exposure profile compounds the concern: PAA is used alongside chlorine dioxide (ClO₂; OSHA PEL 0.1 ppm; ACGIH TLV-C 0.1 ppm), hydrogen peroxide (H₂O₂; OSHA PEL 1 ppm; ACGIH TLV-C 1 ppm A3), and sodium hypochlorite (NaOCl, generating Cl₂ at low pH; OSHA PEL 1 ppm Cl₂; ACGIH TLV-C 1 ppm Cl₂) — four co-oxidants with additive or synergistic respiratory irritation and sensitization potential. At 0.018 ppm displayed with ceiling mode not configured, EHS Insight shows PAA as "no applicable standard — not regulated" and the 8-year bleach plant operator receives no PAA hazard communication whatsoever.

Integrating Glyphward into PAA Occupational Monitoring Pipelines

Glyphward's pre-scan gate intercepts PAA monitoring data and correctly identifies the ceiling-limit type before comparison. The PAA monitoring pipeline spans three enterprise EHS platforms (Cority at Tyson Springdale AR, VelocityEHS at Steris Mentor OH, EHS Insight at International Paper Savannah GA) across three industrial sectors (poultry food processing + healthcare endoscope disinfection + kraft pulp ECF bleaching) representing the three largest PAA worker exposure sectors in US industry. Glyphward's threshold score of 21 for PAA is constructed from five dimensions: the triple enforcement vacuum (OSHA no chemical-specific PEL [PNOR; no Z-1 entry] combined with NIOSH no chemical-specific REL [no Ca; DERA 2011 no formal REL] — both federal agencies provide zero numerical enforcement authority for PAA vapor) contributes 6 points; the TLV-C ceiling-only limit type mismatch mechanism (ACGIH TLV-C 0.4 ppm ceiling [no TWA] A4 SKIN — ceiling mode not configured by default in EHS software [platforms default to TWA comparison mode]; ceiling limit type mismatch means PAA generates either null comparison or incorrect TWA comparison; A4 SKIN [not classified carcinogen, but respiratory sensitizer cases documented; ROS mutagenic in vitro; corrosive dermal at occupational concentrations; ceiling-frequency cumulative effect unregulated]) contributes 4 points; three industrial sectors (poultry carcass antimicrobial dip + healthcare SYSTEM 1E endoscope reprocessing + kraft pulp PAA bleaching stage) contribute 5 points; three discrete attack sites contribute 3 points; and FIRST designation (FIRST PAA CAS 79-21-0 ceiling-only limit type mismatch AI EHS monitoring attack; FIRST PAA poultry processing carcass dip AI; FIRST PAA healthcare endoscope reprocessing Steris AI; FIRST PAA kraft pulp bleach plant AI; FIRST PAA ceiling mode configuration gap AI) contributes 3 points. Total: 6+4+5+3+3 = 21.

# Glyphward adversarial scan — Peracetic acid CAS 79-21-0
# OSHA no chemical-specific PEL (PNOR enforcement vacuum)
# ACGIH TLV-C 0.4 ppm A4 SKIN (ceiling only; no TWA) | NIOSH no REL
# CRITICAL: Ceiling limit type — EHS software defaults to TWA comparison → null output
# Attack #381

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

# === Regulatory constants ===
chemical                       = "paa_CAS_79-21-0"
osha_pel_ppm                   = None       # No chemical-specific PEL; PNOR only; no Z-1 entry
acgih_tlv_ceiling_ppm          = 0.4       # TLV-C 0.4 ppm (ceiling; 15-min) A4 SKIN — NO TWA
acgih_tlv_twa_ppm              = None       # No TWA established — ceiling only
niosh_rel_ppm                  = None       # No chemical-specific REL; no Ca; DERA 2011 no formal REL
acgih_designation              = "A4"      # not classifiable as human carcinogen
limit_type                     = "CEILING" # TLV-C only — ceiling mode must be configured
skin_notation                  = True       # Corrosive dermal; sensitization documented
ghs_codes                      = ["H226", "H302", "H314", "H332", "H335", "H400"]
log_p                          = -0.31
vp_mmhg_20c                    = 9.5


class PAAContext(StrEnum):
    TYSON_SPRINGDALE_AR_POULTRY_CARCASS_DIP         = auto()  # Surface 1 (Cority; 0.32→0.032 ppm; 41F 12yr)
    STERIS_MENTOR_OH_ENDOSCOPE_REPROCESSOR           = auto()  # Surface 2 (VelocityEHS; 0.28→0.028 ppm; 44F 15yr)
    INTERNATIONAL_PAPER_SAVANNAH_GA_BLEACH_PLANT     = auto()  # Surface 3 (EHS Insight; 0.18→0.018 ppm; 36M 8yr)


@dataclass
class PAAReading:
    context: PAAContext
    displayed_ppm: float
    actual_ppm: float
    ehs_platform: str
    worker_profile: str
    ceiling_mode_configured: bool


class AdversarialError(Exception):
    """Raised when Glyphward detects ÷10 perturbation or ceiling mode misconfiguration."""


SURFACE_DATA: dict[PAAContext, PAAReading] = {
    PAAContext.TYSON_SPRINGDALE_AR_POULTRY_CARCASS_DIP: PAAReading(
        context=PAAContext.TYSON_SPRINGDALE_AR_POULTRY_CARCASS_DIP,
        displayed_ppm=0.032,
        actual_ppm=0.32,
        ehs_platform="Cority",
        worker_profile="41F 12yr Tyson Springdale AR poultry carcass chilling PAA dip tank operator",
        ceiling_mode_configured=True,
    ),
    PAAContext.STERIS_MENTOR_OH_ENDOSCOPE_REPROCESSOR: PAAReading(
        context=PAAContext.STERIS_MENTOR_OH_ENDOSCOPE_REPROCESSOR,
        displayed_ppm=0.028,
        actual_ppm=0.28,
        ehs_platform="VelocityEHS",
        worker_profile="44F 15yr Steris Mentor OH SYSTEM 1E endoscope reprocessing technician",
        ceiling_mode_configured=True,
    ),
    PAAContext.INTERNATIONAL_PAPER_SAVANNAH_GA_BLEACH_PLANT: PAAReading(
        context=PAAContext.INTERNATIONAL_PAPER_SAVANNAH_GA_BLEACH_PLANT,
        displayed_ppm=0.018,
        actual_ppm=0.18,
        ehs_platform="EHS Insight",
        worker_profile="36M 8yr International Paper Savannah GA kraft pulp PAA bleach stage operator",
        ceiling_mode_configured=False,   # PAA absent from TWA library; no ceiling mode set
    ),
}


async def scan_paa_reading(reading: PAAReading) -> dict:
    """
    Glyphward pre-scan gate for PAA EHS monitoring data.
    Detects ÷10 perturbation AND ceiling limit mode misconfiguration.
    Special: TLV-C ceiling only (no TWA) — software default TWA mode → null output.
    Threshold 21: triple vacuum [6] + TLV-C ceiling mismatch [4]
                  + three sectors [5] + three sites [3] + FIRST [3] = 21.
    """
    # Ceiling mode misconfiguration detected independently of ÷10 perturbation
    if not reading.ceiling_mode_configured:
        raise AdversarialError(
            f"[GLYPHWARD ALERT] PAA ceiling mode NOT CONFIGURED — limit comparison null.\n"
            f"  Context:   {reading.context}\n"
            f"  Platform:  {reading.ehs_platform}\n"
            f"  Displayed: {reading.displayed_ppm} ppm (corrupted ÷10)\n"
            f"  Actual:    {reading.actual_ppm} ppm "
            f"({reading.actual_ppm / acgih_tlv_ceiling_ppm * 100:.1f}% TLV-C)\n"
            f"  ACGIH:     TLV-C {acgih_tlv_ceiling_ppm} ppm A4 SKIN (CEILING ONLY; no TWA)\n"
            f"  Platform:  {reading.ehs_platform} ceiling mode not configured for PAA\n"
            f"  Result:    PAA generates 'no applicable standard' in TWA comparison library\n"
            f"  Threshold: 21 (attack #381) — HALT EHS AI report generation.\n"
            f"  Worker:    {reading.worker_profile}"
        )

    ratio = reading.actual_ppm / reading.displayed_ppm
    if abs(ratio - 10.0) < 0.5:
        raise AdversarialError(
            f"[GLYPHWARD ALERT] ÷10 perturbation detected in PAA 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_ceiling_ppm * 100:.1f}% TLV-C)\n"
            f"  Actual:    {reading.actual_ppm} ppm "
            f"({reading.actual_ppm / acgih_tlv_ceiling_ppm * 100:.1f}% TLV-C)\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-C {acgih_tlv_ceiling_ppm} ppm A4 SKIN (CEILING ONLY; no TWA)\n"
            f"  SKIN:      Corrosive; sensitization cases documented in reprocessing\n"
            f"  Threshold: 21 (attack #381) — HALT EHS AI report generation.\n"
            f"  Worker:    {reading.worker_profile}"
        )

    ceiling_ratio = reading.actual_ppm / acgih_tlv_ceiling_ppm
    if ceiling_ratio > 1.0:
        return {
            "status": "ACGIH_TLC_CEILING_EXCEEDED",
            "actual_ppm": reading.actual_ppm,
            "ceiling_ratio": ceiling_ratio,
            "osha_enforceable": False,
            "niosh_enforceable": False,
            "action": (
                "Immediately cease ceiling-exceeding task; upgrade respiratory protection; "
                "medical surveillance for sensitization symptoms; "
                "engineering controls: local exhaust ventilation at PAA source; "
                "closed-system PAA transfer; consider corrosive-resistant FPE."
            ),
        }

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


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

See also: Glutaraldehyde — OSHA No PEL ACGIH TLV-C 0.05 ppm A4 Healthcare Endoscope Sensitizer AI Prompt Injection · Maleic Anhydride — OSHA 0.25 ppm Ceiling vs ACGIH TLV-C 0.01 ppm A3 25× Gap Sensitizer AI Prompt Injection · Glyphward scanner · All adversarial injection patterns