Automated testing and assurance tooling for CASTÚO-SYSTEM™. Performance benchmarking is a
TARGET, not a current capability — see §4.
License: pending IP review. This repository is not open source: no license has been granted. Until a
LICENSEfile is published, default copyright applies (all rights reserved by the author). The previous AGPL-3.0 badge was removed because no license file backed it and no licensing decision has been taken.Contributions: external contributions are not accepted until a contributor licence agreement (CLA) or contribution policy is defined.
Cast-o is the quality assurance and performance engine of the ecosystem. It provides a unified environment to structure, automate, and validate the CASTÚO-SYSTEM ecosystem through tests, infrastructure-as-code, and integration tools.
Its scope covers:
- Automated Testing: Unit and integration tests (
pytest). - Infrastructure Validation: Terraform (Hetzner) and Kubernetes manifests.
- IoT & Edge Simulation: ESP32 code and MQTT integration testing.
- Performance Benchmarking (
TARGET): Regression detection and resource usage analysis — not implemented yet. - CI/CD Support: Reusable pipelines and hardening checklists.
Cast-o acts as the TOOLING anchor, providing the necessary infrastructure for technical validation across all layers.
Cast-o (Assurance)
│
├── castuo-evidence (Public Fabric)
│ Evidence verification target
│
├── CASTÚO-SYSTEM (Private Core)
│ Execution engine
│
└── castuo-evolution (External surface)
Prepared external surface — not the current
GitHub authority nor a synchronised SSOT
Canonical authority: Castuo-system (private) is the current authority for code, operational documentation and technical evolution. castuo-evolution is an external, prepared surface; it is not presented as the current GitHub authority or a synchronised source of truth.
- Testing Framework: Unit, integration, and E2E tests for AI and IoT components.
- Dockerized Environments: Modular
docker-composefiles for IoT, Cloud, and HA scenarios. - Infrastructure as Code: Terraform assets for Hetzner and K8s manifests.
- Observability: Monitoring configurations for Prometheus and Grafana.
- Documentation & Diagnostics: System diagnostics, integration checklists, and contingency reports.
Every claim uses one status: CURRENT (implemented and verifiable) · TARGET (approved objective, not implemented) · EXPERIMENTAL (prototype, not consolidated) · PENDING (planned, or evidence incomplete) · NOT_CLAIMED (not implemented; not presented as a capability).
| Claim | Status | Evidence / limitation |
|---|---|---|
Unit and integration test suite (pytest) |
CURRENT (partial) |
The suite is only partly executable. Last documented run — commit d57936c, 2026-09-28, local Windows, Python 3.12: 463 passed, 3 failed, 7 errors, and 1 file could not be loaded (tests/test_router_hardening.py). Stabilising the suite is pending. The Python validation workflow fails on main (flake8 E999: test_e2e.py is a bash script). |
| CI/CD automation (GitHub Actions) | CURRENT (partial) |
Several workflows run on schedule. Known reds on main: Agent Sync Hardening (drift false positive, fix in PR #27), Reconcile CI/CD, 48h Operativity Gate, docker-security. |
| Docker Compose environments | PENDING |
Compose files exist (IoT, cloud, HA, …); no recorded run evidence. |
| Infrastructure as Code (Terraform, Kubernetes) | PENDING |
3 .tf files and K8s manifests exist; no plan/apply evidence. |
| Observability (Prometheus, Grafana) | PENDING |
Configuration files exist; no deployment evidence from this repository. |
| IoT / ESP32 / MQTT simulation | EXPERIMENTAL |
Sample code present; not consolidated or tested end to end. |
| Performance benchmarking | TARGET |
No benchmark code, dataset, method or results in this repository. Per docs/CASTUO_ARCHITECTURE_GOVERNANCE.md, a benchmark is valid only with dataset, version, environment and method. |
| E2E browser tests (Playwright) | NOT_CLAIMED |
Playwright is not present in this repository. |
| Test records federated into CASTÚO-EVOLUTION | NOT_CLAIMED |
No federation mechanism is implemented. |
git clone https://github.com/Traky12/Cast-o.git
cd Cast-o
cp .env.example .env
pytest tests/ -v # partial: see §4 for current failuresdocker compose up -d is PENDING verification (see §4).
← Profile | → Evidence | → Governance scope | → Architecture Docs
Build · Validate · Observe · Document · Evolve
This repository is governed through the CASTÚO-SYSTEM evidence chain. Its current role, visibility boundary, required provenance, security baseline and promotion rules are defined in docs/CASTUO_ARCHITECTURE_GOVERNANCE.md. A repository artifact or green workflow proves only the declared scope; it does not by itself prove certification, production operation, funding, customer contracts or commercial success.
The assurance suite must test both accepted and rejected paths: unregistered agent, unauthorised tool, incorrect tenant, revoked credential, duplicate replay, missing evidence hash, unapproved model, sensitive action without approval, disconnected node, synchronisation conflict, rollback request and incompatible version.
The minimum failure contract is:
denied request → logged → explainable → recoverable
A passing local test proves only the declared test scope. It does not prove federated operation, production security, customer adoption or regulatory conformity.
This repository is part of the CASTÚO-SYSTEM private-cloud target architecture. Its repository scope does not by itself prove cloud provisioning, DNS, production operation, customer traction, financing, certification or independent validation. The service identity is a governed target boundary until a deployment record, access control, health check, observability, backup, restore, rollback, owner and dated Evidence Center record are published.
Public claims use the status taxonomy in §4. OpenClaw and n8n, where referenced, are optional third-party compatibility adapters and not the sovereign governance control plane.