cs-stratcore

cs-stratcore is the sovereign strategic interaction subsystem in the CounterSec / SAVANT ecosystem.

This repository has completed bounded implementation through Pass 5:

• doctrine and scaffold,
• executable foundation,
• canonical schemas,
• bounded strategic assessment API flow,
• explicit external adapter seams (deferred, not integrated).

Mission

cs-stratcore provides bounded strategic assessment over governed state, including actor-model framing, objective-conflict analysis, move/countermove estimation, scenario scoring, escalation-aware risk framing, and recommendation candidate generation.

Design Doctrine

• Not a generic game-theory playground.
• Not a truth authority.
• Does not use raw uncontrolled text as the primary reasoning substrate.
• Consumes curated structured state from sovereign upstream systems.
• Emits structured, explainable, auditable strategic assessments and recommendation candidates.

Repository Scope (Current State)

• Single Python project with src/ layout.
• Unified namespace: cs_stratcore.
• FastAPI app entrypoint at src/cs_stratcore/api/main.py with:
  • GET /health returning {"service": "cs-stratcore", "status": "ok"}
  • deterministic POST /strategic-assessment
• Canonical internal schemas across actor, scenario, belief, strategy, recommendation, and scoring/audit domains.
• Deterministic placeholder service orchestration for strategic assessment.
• Adapter seam contracts for OpenSpiel, POSGGym, and attack-graphs with deferred stubs (NotImplementedError).
• CI workflow for lint, type-check, and tests.

No external runtime integration is included yet for OpenSpiel, POSGGym, or attack-graphs.

Planned Ecosystem Relationships

cs-stratcore is designed to operate alongside:

• cs-darkweb (collection operations)
• CounterSecAI (AI-native operational interaction)
• engraphify (governed relational intelligence / graph state)
• cs-mirocore (environment and scenario-state construction)
• ASH (assurance, policy, audit, replay, promotion)
• SAVANT (mission-level federated operating system)

Adapter seams are present for:

• OpenSpiel
• POSGGym
• attack-graphs

Structure

cs-stratcore/
  .github/workflows/ci.yml
  docs/architecture/
    overview.md
    module_map.md
    rules.md
  src/cs_stratcore/
    api/
    common/
    actor_models/
    scenario_engine/
    payoff_engine/
    strategy_engine/
    belief_engine/
    recommender/
    scoring_audit/
    adapters/
      openspiel/
      posggym/
      attack_graphs/
  tests/
  pyproject.toml

Status

cs-stratcore is structurally complete through Pass 5. The architecture is stable, canonical model ownership remains internal, service behavior is deterministic and bounded, and external capability seams are explicit but deferred.