ASI Backbone Learning

Practical .NET architecture education for governed execution, secure applications, AI integration, and policy-driven systems.

AsiBackbone/Learning teaches how proposed operations can move through explicit policy decisions, acknowledgments, scoped authority, host-owned execution, and durable audit evidence.

You can use this material without installing the AsiBackbone framework. Tutorials, samples, comparisons, and labs are intended to remain useful as independent architecture education.

In this project, ASI means Accountable Systems Infrastructure. Learning is the educational layer of the ASI Backbone organization; it is not an artificial general intelligence or artificial superintelligence implementation.

Quick Start — Run It in 10 Minutes

Prefer to see the architecture run before reading the deeper explanation? The foundational Decision Before Execution sample provides the shortest path from clone to observable behavior.

Prerequisite: .NET 10 SDK

Learning selects Microsoft.Testing.Platform in global.json, and every sample test project uses the aligned xunit.v3 runner integration. The same dotnet test commands are used locally and in CI.

From a terminal:

git clone https://github.com/AsiBackbone/Learning.git
cd Learning

dotnet run --project samples/decision-before-execution/Sample/DecisionBeforeExecution.csproj

dotnet test samples/decision-before-execution/Tests/DecisionBeforeExecution.Tests.csproj

Windows path-length troubleshooting

The repository is intended to clone and build normally on supported Windows environments.

If Git reports Filename too long during clone or checkout, you can enable Git for Windows long-path support:

git config --global core.longpaths true

You can also reduce total path length by cloning the repository into a short location such as:

C:\src\Learning

What to Observe

The sample makes one architectural invariant visible:

A blocked decision never reaches the executor.

Allowed decision
   ↓
Host-owned executor invoked

Denied / deferred / escalation-recommended / acknowledgment-required decision
   ↓
Executor not invoked

The console sample evaluates five deterministic scenarios and verifies that exactly one allowed operation crosses the execution boundary. The focused xUnit tests make the same contract repeatable for local development and CI.

Want to understand why this boundary exists or experiment with it?

• Decision Before Execution sample README
• Decision Before Execution tutorial
• Getting Started

Choose Your Next Step

If you want to...	Start here
Learn the foundational governed-execution boundary	Decision Before Execution
Route from a problem you already recognize	Find Your Path
See the curriculum and prerequisites at a glance	Learning Path Map
Decide whether ASP.NET Core authorization is already enough	When ASP.NET Core Authorization Is Enough

What This Architecture Looks Like in Practice

A governed system makes the proposed operation, relevant context, active constraint, decision, and execution boundary visible.

Proposed operation:
account.disable

Context:
Account is protected

Constraint:
Protected accounts require escalation

Decision:
EscalationRecommended

Execution:
Not invoked

Learning explores the architectural boundaries that make behavior like this explicit, testable, auditable, and separable from execution. The same reasoning can be applied to administrative workflows, infrastructure changes, sensitive API operations, AI tool calls, and other consequential actions.

Foundational Learning Path

The established five-part sequence moves from proposed intent to governed AI-assisted execution:

1. Decision Before Execution
2. Policy Context and Explicit Decision Outcomes
3. Acknowledgment and Audit Residue
4. Scoped Capability and Host-Owned Execution
5. Governed AI Tool Gateway

Each foundational topic is reinforced by runnable samples, focused architectural-invariant tests, and hands-on labs.

Want to understand why Learning uses a problem-first tutorial model, how tutorials differ from labs, or how canonical and alternative patterns are handled? See the Learning Model.

ASI Backbone Ecosystem

The organization contains complementary projects with different responsibilities:

Project	Primary role
Learning	Teaches concepts, patterns, tradeoffs, samples, and labs
AsiBackbone	Working governance and policy-control implementation
NetCoreApplicationTemplate	ASP.NET Core reference application and architecture specimen

Learning connects to the implementation repositories when fuller examples are useful, but adoption of either implementation repository is not required to benefit from the educational material.

Scope and Boundaries

ASI Backbone Learning is an educational and architectural resource.

• It teaches architectural patterns; it does not certify compliance or guarantee security.
• Examples do not replace application-specific security, legal, regulatory, safety, or operational review.
• Learning is not an AI model, an artificial general intelligence or artificial superintelligence implementation, or a robotics controller.
• No AsiBackbone package is required, and no pattern is presented as universally correct.

Production systems remain responsible for their own authentication, authorization, persistence, infrastructure, threat modeling, safety controls, regulatory requirements, and operational execution.

Project and Community

Use the canonical project surfaces for deeper information rather than treating the root README as the full reference manual:

Questions and ideas → Discussions · Concrete work → Issues · Changes → Pull Requests

• Published documentation: asibackbone.github.io/Learning
• XML sitemap: https://asibackbone.github.io/Learning/sitemap.xml — submit this canonical sitemap URL to Google Search Console and Bing Webmaster Tools.
• Contribution guidance: CONTRIBUTING.md
• Project status and planned work: ROADMAP.md
• Learning discussions: AsiBackbone/Learning Discussions
• Organization-wide discussion: ASI Backbone Organization Discussions
• Governance: GOVERNANCE.md
• Security policy: SECURITY.md
• Citation metadata: CITATION.cff
• Licensing details: LICENSING.md
• X publication operations: X_PUBLISHING.md

Search discovery and change notification

The XML sitemap is the complete canonical discovery inventory for the published Learning site. IndexNow complements that inventory with a narrower post-deployment signal: after GitHub Pages deploys successfully, the publication workflow notifies participating search engines only about canonical Learning URLs that were added, modified, or removed since the previously deployed revision. IndexNow does not replace the sitemap and does not determine whether documentation is publishable.

The site publishes deployment-revision.txt so the next publication can compare against the revision actually represented by GitHub Pages rather than assuming the immediately preceding commit was deployed. On the first run, the workflow falls back to the push base revision. The public indexnow-key.txt file verifies control of the /Learning/ URL space through IndexNow keyLocation; it is intentionally public protocol verification material, not a private workflow credential. IndexNow service failures are reported as non-blocking warnings after deployment.

An optional, separate post-deployment publisher can announce newly eligible feed: true documents on X. It consumes the same source frontmatter as RSS but does not read or modify feed.xml. Durable receipts, canonical-URL reconciliation, and a deployed-revision cursor provide retryable, effectively-once delivery without making X availability a documentation deployment dependency. See the X Publication Runbook for the security boundary, setup, dry-run, retry, recovery, and disable procedures.

Issues are best used for concrete repository work; Learning Discussions are better suited to exploratory architecture questions, tutorial proposals, alternatives, design debates, and community examples.

Project Status

Active development — foundational tutorial, sample, test, and lab path established.

Current development is focused on stronger implementation references, deeper labs, ASP.NET Core architecture, security and trust architecture, governance material, architecture comparisons, and improved discoverability. See ROADMAP.md for the maintained direction.

Citing a Release

Use the Zenodo concept DOI when citing the evolving Learning work as a whole.

When reproducibility depends on the exact material reviewed, cite the version-specific DOI shown on that release's Zenodo record and include the corresponding Learning version or GitHub tag. GitHub Releases provides the versioned repository trail.

Stable GitHub Releases also attach durable, anonymously downloadable release notes, a scoped SPDX sample/dependency inventory, and a SHA-256 evidence manifest. See the Stable Release Evidence Runbook for scope and verification commands.

License

ASI Backbone Learning uses component-specific licensing:

• Documentation, educational material, and diagrams: CC BY 4.0
• Executable sample code under samples/: MIT License
• Source-code snippets embedded in documentation: MIT License unless otherwise noted

See LICENSING.md for the complete licensing policy.

---

ASI Backbone Learning is not intended to provide doctrine. It is intended to provide patterns worth examining.

Read them. Test them. Challenge them. Adapt them. Improve them.