Triage a process or chemical inventory for safety hazards by screening reactive-chemistry incompatibilities, flagging hazardous properties (flammability, toxicity, thermal instability), and summarizing key risks using Paramus data and property tools.When to use: a user asks “what are the safety risks of this process/mixture?”.Multi-tool WORKFLOW with process-safety judgment.
Paramus — Process Safety Triage
Overview
Screen an inventory/process for hazardous properties and reactive incompatibilities, then summarize the dominant risks. The value is the fused hazard picture + safety judgment, not one property lookup.
Guidance vs execution: hazard properties and compatibility come from tested Paramus data/tools. Do not assert safety conclusions from model memory — this is a domain where errors are dangerous.
When to Use
- Early hazard screening of a process, mixture, or chemical inventory
- Flagging incompatible-storage / reactive-hazard combinations Do not use for: formal HAZOP/LOPA or regulatory safety cases (those require qualified human review).
Workflow
Chemical inventory / process streams + conditions
↓ 1. Resolve components → identifiers + hazard data lookup
↓ 2. Hazardous-property flags → flammability, toxicity, thermal instability, corrosivity
↓ 3. Reactive incompatibility check → pairwise incompatibility matrix
↓ 4. Condition-aware weighting → T/P amplifying a hazard
↓ 5. Risk summary & report → ranked hazards + incompatibilities + provenance + disclaimer
Procedure
- Discover tools via
search/get_schema("hazard properties", "chemical compatibility"). - Resolve components; retrieve hazard data with provenance.
- Flag hazardous properties per component.
- Screen pairwise reactive incompatibilities across the inventory.
- Weight by process conditions; report ranked risks, incompatibilities, and a clear scope disclaimer.
Domain Judgment
- This is a screen, not a safety assessment — state prominently that qualified human review is required.
- Absence of a flag ≠ safe; state data-coverage limits explicitly.
- Incompatibility hits are high-priority — surface them first, never bury them in a score.
Fallbacks
- Hazard data missing for a component → mark it "unassessed"; never imply it is safe.
- Endpoint unreachable → stop and report; do not infer hazards from local heuristics.
Tools this skill may use
Candidate deterministic tools an agent is likely to route to when running this skill. The skill decides which to call at runtime.
-
Calculate Flash Point
ThermodynamicsCalculate critical safety properties including flash point, autoignition temperature, lower flammability limit (LFL), and upper flammability limit (UFL). Essential for process safety, hazard assessment, and storage/handling requirements of flammable chemicals. -
Get Safety Data
ThermodynamicsUser needs safety information for handling, storage, or regulatory compliance for methanol. -
Stockroom Get Hazard Info
Lab ManagementGet GHS pictograms, H-statements, and ADR transport classification for a stockroom chemical. Returns a deep_link URL pointing to the matching view in the app UI. -
Compliance Classify Mixture
OtherPreliminary CLP mixture classification from component weight % and ECHA harmonised/self-classifications. Screening only - not a binding determination. -
Cantera Ignition Delay
Chemical KineticsCalculate the ignition delay time for a fuel/oxidizer mixture in a batch reactor. The ignition delay is the time required for autoignition to occur, characterized by a rapid temperature rise. This is critical for understanding engine knock, combustion timing, and safety.
Browse the full deterministic layer in the tool browser.
Used in these use cases
Customer scenarios that orchestrate this skill end-to-end.