Warfighting Agent Suite
Doctrinally grounded military AI agents. One sovereign mesh.
Each agent is purpose-built around a doctrinal task, bounded by policy, instrumented for assurance, and orchestrated through the sovereign Agent Mesh. Every SynapseCommand agent is purpose-built for a doctrinal task and validated against the same verification programme. Agents reason in parallel and coordinate through the sovereign Agent Mesh, with the combination set by mission requirements.
Fourteen agents. Eight are stated at TRL 7 and six have TRL 7 demonstration pending. These are first-party statements about the private platform; the public integration repository does not establish them. Thirteen agents have a recorded verification result at the 27 August 2026 baseline; SC-MET-01 Weather is assigned to the programme and has no recorded result yet.
Agent SC-RED-01
Red-Force Behaviour Agent
Adversary doctrine simulation & RL red-teaming.
Purpose
Commanders need to understand what an adversary can do before they commit friendly forces, but generic threat summaries hide the doctrine that shapes real behaviour. SC-RED-01 models adversary intent and likely courses of action by combining reinforcement-learned policies with retrieval over sovereign doctrinal corpora. It conditions behaviour on observed pattern-of-life and ISR indicators, then returns probable COAs, most-dangerous and most-likely postures with explicit confidence intervals. Every inference is tagged to the source doctrine and sensor data that shaped it, so red-teaming outputs can be audited rather than trusted.
Doctrine alignment
Designed to support JIPOE workflows described in STANAG 6527 / AIntP-17.
Typical inputs
- Adversary OOB
- Historical doctrine
- ISR indicators
- Pattern-of-life
Outputs
- Probable adversary COAs
- Most-dangerous / most-likely posture
- Doctrinal red-teaming reports
Underlying products
SynapseMesh™VerityGridAI™
Deployment notes
Sovereign cloud or on-prem; fully air-gappable.
All agents operate under VerityGridAI™ assurance, writing hash-chained, append-only decision records to CyberFort Blockchain™ that are ML-DSA-signed in assured deployments; records produced without an active signing key are marked UNVERIFIED.
Agent SC-BLUE-01
Blue-Force Posture Agent
Friendly readiness, logistics, ISR coverage.
Purpose
Readiness reports often arrive late, conflict across systems, and leave commanders guessing whether a logistics shortfall or sensor gap is real. SC-BLUE-01 maintains a continuously updated picture of friendly readiness, logistics chains and ISR coverage gaps by ingesting force tracker feeds, logistics ERP data and maintenance status. It computes readiness scores, sensor coverage geometry and sustainment risk as independently auditable values, so a commander can see why a score moved, not just that it moved. Every posture snapshot is hashed and anchored to CyberFort Blockchain, giving a tamper-evident history that survives staff turnover and system swaps.
Doctrine alignment
NATO Joint Logistics doctrine and AJP-4 alignment.
Typical inputs
- Force tracker feeds
- Logistics ERP
- ISR plan
- Maintenance status
Outputs
- Readiness score
- Coverage heat-map
- Sustainment risk alerts
Underlying products
SynapseMesh™CyberFort Blockchain™
Deployment notes
Operates within national accredited enclaves; fully air-gappable.
All agents operate under VerityGridAI™ assurance, writing hash-chained, append-only decision records to CyberFort Blockchain™ that are ML-DSA-signed in assured deployments; records produced without an active signing key are marked UNVERIFIED.
Agent SC-ISR-01
ISR Fusion Agent
Multi-source fusion: radar, FMV, SAR, SIGINT, SATINT, AIS, OSINT.
Purpose
The modern operations room is flooded with radar, imagery, signals and open-source feeds, yet fusion systems often average away the very disagreements that warn of deception or uncertainty. SC-ISR-01 fuses radar, FMV, SAR, SIGINT, SATINT, AIS and OSINT into a coherent operational picture with calibrated confidence. The core keeps disagreement between sources visible rather than smoothing it away, and every fused track carries sensor provenance back to its contributing inputs. It runs at the edge under DDIL conditions and reconciles with the core when connectivity returns, so the picture persists even when the network does not.
Doctrine alignment
STANAG 4607 and 4609 data models supported; Link-family inputs via authorised gateways.
Typical inputs
- Radar tracks
- FMV / SAR imagery
- SIGINT, SATINT, AIS
- OSINT streams
Outputs
- Fused tracks with provenance
- Confidence intervals
- Disagreement diagnostics
Underlying products
SynapseMesh™SecureSat Guardian™CryptoSleuth Pro™
Deployment notes
Edge or core; supports DDIL conditions.
All agents operate under VerityGridAI™ assurance, writing hash-chained, append-only decision records to CyberFort Blockchain™ that are ML-DSA-signed in assured deployments; records produced without an active signing key are marked UNVERIFIED.
Agent SC-SCN-01
Scenario Synthesis Agent
Automated scene-setting & dynamic events.
Purpose
Exercise and planning staffs spend weeks hand-building scenarios that drift between iterations and cannot be reproduced for after-action review. SC-SCN-01 generates training, planning and what-if scenarios from a vignette or commander prompt, building scene-setting, plausibility-constrained events and branch-and-sequel trees in checkpointed passes. Each generated package is written to the hash-chained decision record, ML-DSA-signed in assured deployments, and reproducible, so an exercise used today can be replayed exactly years later. Output is confined to training and planning contexts, and the agent never injects events into live operational systems.
Doctrine alignment
Supports NATO Joint Warfare Centre and national LVC training models.
Typical inputs
- Commander intent
- Geographic AOR
- Force composition
Outputs
- Scenario packages
- Dynamic event injects
- Branch & sequel trees
Underlying products
SynapseMesh™VerityGridAI™
Deployment notes
Standalone or coupled to existing simulation backbones; fully air-gappable.
All agents operate under VerityGridAI™ assurance, writing hash-chained, append-only decision records to CyberFort Blockchain™ that are ML-DSA-signed in assured deployments; records produced without an active signing key are marked UNVERIFIED.
Agent SC-COA-01
COA Analysis Engine
RL + Bayesian + Monte Carlo course-of-action exploration.
Purpose
Course-of-action comparison too often rests on a single staff estimate and a spreadsheet, leaving commanders without a clear view of where assumptions break. SC-COA-01 generates and evaluates courses of action across a high-dimensional outcome space using reinforcement learning, Bayesian causal networks and Monte Carlo simulation in parallel. Agreement between methods raises confidence; disagreement is reported as a finding in its own right. Ranking criteria are explicit and commander-adjustable, source records stay at the originating node, and the complete run is sealed to CyberFort Blockchain so any ranking can be re-derived later.
Doctrine alignment
NATO COPD / Comprehensive Operations Planning Directive aligned.
Typical inputs
- Mission objectives
- Constraints & ROE
- Red & Blue postures
Outputs
- Ranked COAs
- Risk distributions
- Sensitivity analysis
Underlying products
SynapseMesh™QuantumProof Protocol™
Deployment notes
Scales horizontally; supports federated runs across allied nodes.
All agents operate under VerityGridAI™ assurance, writing hash-chained, append-only decision records to CyberFort Blockchain™ that are ML-DSA-signed in assured deployments; records produced without an active signing key are marked UNVERIFIED.
Agent SC-TGT-01
Targeting & Engagement Agent
D3A / F2T2EA workflows with deterministic, authenticated ROE policy checks.
Purpose
Targeting decisions carry legal, political and strategic consequences, so decision support must be transparent, deterministic and always subordinate to human authority. SC-TGT-01 provides decision support across the full targeting cycle. It runs D3A and F2T2EA workflows as explicit state machines, applies deterministic ROE compliance checks against authenticated, version-pinned rule sets, and requires recorded human-in-the-loop authorisation at every irreversible gate, signed with ML-DSA in assured deployments. There is no configuration that removes the human gate, and every compliance check leaves a proof that can be inspected after the event.
Doctrine alignment
NATO AJP-3.9 Joint Targeting; D3A and F2T2EA workflows.
Typical inputs
- Target nominations
- ROE policy
- Collateral models
Outputs
- Engagement options with ROE compliance proofs
- Effects assessments
Underlying products
CipherLayer™CyberFort Blockchain™PQ-STRAT™
Deployment notes
Always paired with human-in-the-loop authority gates.
All agents operate under VerityGridAI™ assurance, writing hash-chained, append-only decision records to CyberFort Blockchain™ that are ML-DSA-signed in assured deployments; records produced without an active signing key are marked UNVERIFIED.
Agent SC-PLAN-01
Operational Planning Agent
NATO OPP decision support & order development.
Purpose
Operational planning produces vast staff products, yet drafts often lose traceability to the COAs, estimates and doctrine that produced them. SC-PLAN-01 supports the operational planning process from mission analysis through COA development to OPORD generation. Drafting is retrieval-grounded in the sovereign doctrinal corpus and in outputs from the COA and logic agents, so every annex and synchronisation matrix can be traced back to its source. Coalition-aware federation lets allied cells contribute without raw staff products crossing the boundary, and every draft, revision and final order is versioned and sealed to CyberFort Blockchain.
Doctrine alignment
NATO OPP / AJP-5; national equivalents.
Typical inputs
- Mission orders
- Commander's guidance
- Staff estimates
Outputs
- Annexes & OPORDs
- Synchronisation matrices
- Decision support templates
Underlying products
SynapseMesh™CyberFort Blockchain™
Deployment notes
Sovereign cloud, on-prem classified or air-gapped; coalition-aware federation when connected.
All agents operate under VerityGridAI™ assurance, writing hash-chained, append-only decision records to CyberFort Blockchain™ that are ML-DSA-signed in assured deployments; records produced without an active signing key are marked UNVERIFIED.
Agent SC-LES-01
Lessons-Learned & Assessment Agent
Predicted vs. actual, auditable artefacts.
Purpose
After-action reviews frequently rely on reconstructed narratives and disputed recollections, because the original predictions were never recorded or time-stamped. SC-LES-01 compares predicted to actual outcomes by reading the sealed decision record. Because every prediction is retained in the hash-chained record and, in assured deployments, ML-DSA-signed when produced, variance analysis is an audit rather than a reconstruction. It produces trend analyses and cryptographic audit packages that a third party can verify without trusting the platform, supporting the NATO Lessons Learned Capability framework with evidence that cannot be quietly rewritten.
Doctrine alignment
NATO Lessons Learned Capability framework.
Typical inputs
- Operational logs
- Sensor truth data
- Decision logs
Outputs
- Variance reports
- Trend analyses
- Cryptographic audit packages
Underlying products
CyberFort Blockchain™VerityGridAI™
Deployment notes
Long-term archival on tamper-evident ledger.
All agents operate under VerityGridAI™ assurance, writing hash-chained, append-only decision records to CyberFort Blockchain™ that are ML-DSA-signed in assured deployments; records produced without an active signing key are marked UNVERIFIED.
Agent SC-LOG-01
Logistics Agent
Sustainment and convoy decision support.
Purpose
Sustainment plans fail when assumptions about stock levels, route safety and resupply timing are buried in spreadsheets and out of date before they are read. SC-LOG-01 plans and stress-tests sustainment by modelling stock levels, resupply timing, convoy routing and the effect of contested lines of communication. It generates feasibility assessments and culmination warnings with assumptions stated explicitly, so a shortfall warning can be traced to the input that caused it. The agent tolerates intermittent connectivity by holding state locally, and seals every assessment to CyberFort Blockchain for an auditable supply picture.
Doctrine alignment
NATO AJP-4 logistics doctrine; national sustainment planning practice.
Typical inputs
- Stock and consumption rates
- Movement and transport plans
- Route and threat overlays
- Host-nation support arrangements
Outputs
- Sustainment feasibility assessments
- Convoy routing and timing options
- Culmination and shortfall warnings
Underlying products
SynapseMesh™CyberFort Blockchain™
Deployment notes
Sovereign cloud, on-prem or air-gapped; tolerates intermittent connectivity.
All agents operate under VerityGridAI™ assurance, writing hash-chained, append-only decision records to CyberFort Blockchain™ that are ML-DSA-signed in assured deployments; records produced without an active signing key are marked UNVERIFIED.
Agent SC-LNP-01
Logic & Planning Agent
Structured reasoning and plan decomposition.
Purpose
Staff products can look coherent while containing contradictory tasks, missing dependencies or constraints that silently invalidate the plan. SC-LNP-01 decomposes commander intent into tasks, dependencies and constraints, then checks candidate plans for logical consistency before staff time is spent on them. The core parses a formal task graph, locates contradictions and gaps as specific, located defects, and continuously validates intermediate products from the planning agents. It does not generate plans of its own; it is the sceptic in the mesh, ensuring that what is built is internally consistent and traceable to intent.
Doctrine alignment
Supports mission analysis within NATO OPP / AJP-5.
Typical inputs
- Commander intent
- Task organisation
- Constraints, restraints and assumptions
Outputs
- Task decomposition trees
- Dependency and sequencing checks
- Contradiction and gap reports
Underlying products
SynapseMesh™VerityGridAI™
Deployment notes
Runs alongside the planning agents in the same enclave.
All agents operate under VerityGridAI™ assurance, writing hash-chained, append-only decision records to CyberFort Blockchain™ that are ML-DSA-signed in assured deployments; records produced without an active signing key are marked UNVERIFIED.
Agent SC-STR-01
Strategy & Doctrine Agent
Strategic-level and doctrinal analysis.
Purpose
Operational plans sometimes drift from declared strategy or published doctrine, and the divergence is only discovered after resources are committed. SC-STR-01 relates operational activity to strategic objectives and the doctrinal corpus, surfacing where a plan departs from published doctrine or declared policy. Every assessment cites the passage or instrument it rests on, so a divergence note can be checked in minutes rather than argued for hours. The corpus is versioned, so prior notes remain valid against the doctrine they originally cited, and assessments are sealed to CyberFort Blockchain for a permanent strategic audit trail.
Doctrine alignment
NATO AJP-01 capstone doctrine and national strategic guidance.
Typical inputs
- Strategic guidance and objectives
- Doctrinal publications
- Policy and legal constraints
Outputs
- Doctrine alignment assessments
- Strategic risk and second-order effects
- Policy divergence notes
Underlying products
SynapseMesh™VerityGridAI™
Deployment notes
Headquarters or air-gapped deployment; sovereign corpora only.
All agents operate under VerityGridAI™ assurance, writing hash-chained, append-only decision records to CyberFort Blockchain™ that are ML-DSA-signed in assured deployments; records produced without an active signing key are marked UNVERIFIED.
Agent SC-WHT-01 · Training only
White Force Agent
Exercise control, training variant only.
Training only. Excluded from accredited operational builds and compiled out of operational deployments.
Purpose
Training exercises depend on consistent, fair exercise control, yet human controllers struggle to apply adjudication rules uniformly across long, complex events. SC-WHT-01 acts as exercise control for training and experimentation: it adjudicates injects, tracks training objective coverage and maintains the exercise narrative. Adjudication rules are declared before STARTEX and applied consistently, with every ruling logged alongside its rule and rationale, which ends most post-exercise disputes before they start. The agent is compiled out of operational builds, not merely disabled, so it cannot be accidentally enabled on a live system.
Doctrine alignment
Aligned to NATO exercise planning and Joint Warfare Centre practice.
Typical inputs
- Exercise specification
- Training objectives
- Player actions and reports
Outputs
- Inject adjudication
- Exercise control logs
- Training objective coverage
Underlying products
SynapseMesh™VerityGridAI™
Deployment notes
Training variant only; not fielded on operational systems.
All agents operate under VerityGridAI™ assurance, writing hash-chained, append-only decision records to CyberFort Blockchain™ that are ML-DSA-signed in assured deployments; records produced without an active signing key are marked UNVERIFIED.
Agent SC-ATO-01
Air Tasking Agent
Air tasking order workflows.
Purpose
Air tasking is a high-tempo, multi-constraint problem where late discovery of tanker shortfalls or airspace conflicts can collapse a day's flying programme. SC-ATO-01 supports the air tasking cycle from apportionment to ATO and ACO production, running continuous feasibility checks against available platforms, tanking and airspace. Deconfliction findings are attached to the draft line that causes them, so planners see the conflict at the source, and CipherLayer protects products in transit within the deployment enclave. Every draft, check and revision is sealed to CyberFort Blockchain, giving the air component a complete, auditable tasking history.
Doctrine alignment
NATO AJP-3.3 air and space operations; USAF air tasking cycle.
Typical inputs
- Air apportionment guidance
- Platform and crew availability
- Airspace control measures
- Target nominations
Outputs
- Draft ATO and ACO products
- Sortie feasibility and tanker plans
- Deconfliction findings
Underlying products
SynapseMesh™CipherLayer™
Deployment notes
Delivered as the air-tasking module; sovereign cloud, on-prem classified or air-gapped.
All agents operate under VerityGridAI™ assurance, writing hash-chained, append-only decision records to CyberFort Blockchain™ that are ML-DSA-signed in assured deployments; records produced without an active signing key are marked UNVERIFIED.
Agent SC-MET-01
Weather Agent
Canonical weather representation with deterministic confidence.
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.
Typical inputs
Outputs
Underlying products
SynapseMesh™
Deployment notes
Sovereign cloud, on-prem classified or air-gapped.
All agents operate under VerityGridAI™ assurance, writing hash-chained, append-only decision records to CyberFort Blockchain™ that are ML-DSA-signed in assured deployments; records produced without an active signing key are marked UNVERIFIED.
