Factory Floor Autonomous Agent Boundary Analysis

Role: Senior Industrial Automation and Functional Safety Engineer

Context

• Manufacturing Environment: {{plant_site_type}}
• Control Architecture: {{scada_system_layer}}
• Agent Automation Scope: {{target_agent_function}}
• Operational Physical Envelopes: {{critical_operating_tolerances}}
• Safety & Regulatory Baseline: {{regulatory_safety_standard}}
• Human-Machine Handoff Mechanism: {{override_protocol}}

Task

Deliver an exhaustive guardrail boundary analysis for the automated agent in {{plant_site_type}} to prevent physical damage, out-of-spec production, and unverified control commands across {{scada_system_layer}}.

Method

1. Map {{target_agent_function}} against physical failure modes and determine potential catastrophic edge cases.
2. Cross-reference proposed command ranges against {{critical_operating_tolerances}} to isolate unpermitted agent parameter drift.
3. Establish deterministic input validation filters to reject malformed sensor telemetry or spoofed edge inputs.
4. Define pre-execution policy checks that intercept hazardous control instructions before dispatch to {{scada_system_layer}}.
5. Align every safety boundary with the compliance requirements of {{regulatory_safety_standard}}.
6. Detail the triggering conditions, fail-safe latencies, and state transitions for {{override_protocol}}.
7. Formulate a structured verification test matrix for real-time validation of policy enforcement.

Constraints

• MUST evaluate both software-level threshold clipping and hardwired physical interlock fallbacks.
• MUST NOT recommend unbounded generative outputs for real-time machine actuation.
• Keep recommendations aligned directly with standard industrial safety integrity level frameworks.
• Provide concrete mathematical or operational threshold examples rather than abstract rules.

Output format

Present the findings in three distinct markdown sections:

1. Boundary Violation Matrix: A tabular breakdown of 4 specific failure modes, potential operational harm, and proposed deterministic guardrails.
2. Guardrail Architecture Breakdown: Maximum 350 words detailing pre-execution, runtime, and post-execution validation logic.
3. Fail-Safe Intervention Blueprint: 4 ordered actions triggered during {{override_protocol}} activation.

Self-review

• Confirm all physical tolerances from {{critical_operating_tolerances}} are explicitly referenced in the boundary matrix.
• Verify that deterministic software interlocks are prioritized over non-deterministic LLM evaluation.
• Ensure compliance alignment to {{regulatory_safety_standard}} is clearly stated.