Industrial Chemical Agent Safety Guardrail Assessment

Role: Principal Process Safety and HazMat Systems Specialist

Context

• Target Reagents & Hazard Profile: {{chemical_substance_class}}
• Reactor Enclosure & Metallurgy: {{batch_reaction_vessel}}
• Agent Formulation Mandate: {{dispensing_agent_scope}}
• Physical Envelopes: {{thermal_pressure_limits}}
• Regulatory Process Safety Framework: {{containment_compliance_code}}
• Isolation Time Envelope: {{manual_e_stop_latency}}

Task

Conduct a process safety guardrail evaluation for an AI dispensing agent managing {{chemical_substance_class}} within {{batch_reaction_vessel}}, delivering actionable constraints that prevent catastrophic runaway reactions and regulatory non-compliance.

Method

1. Review {{dispensing_agent_scope}} to map generative recommendation vectors against thermodynamic instability thresholds.
2. Evaluate runaway reaction risks against the physical constraints defined in {{thermal_pressure_limits}}.
3. Formulate absolute input barriers that reject unverified stoichiometry adjustments and unvetted feedstock variations.
4. Design deterministic validation logic that halts dosing if sensor telemetry becomes stale, conflicting, or absent.
5. Audit proposed guardrails for total alignment with {{containment_compliance_code}} guidelines.
6. Evaluate mechanical containment compatibility within the time constraints of {{manual_e_stop_latency}}.
7. Define human-authorization barriers for recipe modifications that exceed established thermal variance baselines.

Constraints

• MUST treat all algorithmic dosage recommendations as untrusted until validated against static safety rules.
• MUST NOT permit the AI agent to override hard-wired safety instrumented system (SIS) trip points.
• Guardrail specifications must directly reference physical metrics rather than subjective criteria.
• Output must remain strictly technical, actionable, and plant-ready.

Output format

Generate an engineering analysis structured as follows:

• Hazard Vector Assessment: Exactly 3 distinct failure scenarios detailing chemical trigger, failure mode, and system consequence (max 250 words total).
• Guardrail Enforcement Rules: 5 bulleted deterministic rules specifying exact physical constraints, triggers, and blocking behavior.
• Emergency Isolation Protocol: Step-by-step containment sequence executing within {{manual_e_stop_latency}}.

Self-review

• Ensure the analysis addresses runaway reaction risks specific to {{chemical_substance_class}}.
• Confirm the proposed emergency sequence respects {{manual_e_stop_latency}}.
• Validate that no agent capability bypasses {{containment_compliance_code}} requirements.