Weather Agent
Canonical weather representation with deterministic confidence.
- Deployment
- Sovereign cloud, on-prem classified or air-gapped.
- Underlying products
- SynapseMesh™
- Stated maturity:
- TRL 7 demonstration pending; no TRL stated
- Verification:
- Assigned to the verification programme; no result recorded yet
- Source:
- First-party statement about the private platform, not established by the public integration repository
Purpose
The single source of weather within SynapseCommand. Rather than ingesting raw meteorological files directly, per-feed adapters normalise national and coalition feeds into one canonical weather record, so downstream agents consume a consistent representation regardless of the originating service. The agent exposes three tools: forecast_at for a point query, forecast_window for a bounded period, and feed_status for the availability and currency of each source. It does not return weather grids. That constraint is deliberate, bounding token count and response latency, and keeping the interface to values a decision actually turns on rather than bulk field data. Confidence is computed deterministically, from feed uncertainty products or published skill-score decay tables, and is never authored by a language model.
Architecture
Per-feed adapters normalise national and coalition meteorological feeds into one canonical weather record, so downstream agents consume a consistent representation regardless of the originating service. The agent exposes three tools to the mesh: forecast_at for a point query, forecast_window for a bounded period, and feed_status for the availability and currency of each source. It does not return weather grids; the interface is bounded to the values a decision turns on, which bounds token count and response latency. Confidence is computed deterministically, from feed uncertainty products or published skill-score decay tables, and is never authored by a language model.
Hover or focus a band to see what it guarantees.
Position in the mesh
- Consumes
- Publishes
- No upstream agent. Source data enters through the agent's own inputs.
- No downstream agent is published. Consumers depend on deployment configuration.
What it does not do
It does not return weather grids, and its confidence values are never authored by a language model.
The same four gates as every agent
An unregistered or unapproved model tag is refused before inference, not flagged afterwards.
Every action is scoped to a declared role. Scope is checked at the engine, not in the prompt.
New or changed models run under supervision until their evaluation evidence is accepted.
Irreversible decisions require a signed human authorisation recorded with the decision.
The agents differ in what they reason about. They do not differ in what they are allowed to do.
Deployment
In-country, accredited cloud, under your jurisdiction and operational control.
Designed for deployment up to NATO SECRET. Accreditation is per deployment.
Runs disconnected. Inference, agents and the decision record are all local.
Ready to see SynapseCommand® in action?
Briefings available under NDA. Classified briefings arranged through cleared channels.