Document intent-aware security verification for non-orthodox code

[epi13]

Outcome

Adds an experimental, non-normative design note for integrating security hardening into the MNCS micro-verifier network without treating unconventional implementation structure as automatic failure.

The proposal:

• separates confirmed invariant violations from suspicious patterns and current exploit-chain reachability;
• establishes the principles that orthodoxy is a heuristic, invariants are authoritative, and declared intent cannot waive failed safety properties;
• defines local, trust-boundary, and composition verification layers;
• proposes machine-readable intentional-deviation records bound to artifact, compiler, target, evidence, and revalidation identities;
• prevents broad syntax whitelists and warning-only suppression;
• outlines an evidence-backed agentic repair loop and recursive learning model that does not homogenize implementations;
• defines staged Forge, micro-verifier, RAVEL, and governance integration boundaries; and
• explicitly preserves existing MNCS, MNCDS, evidence-class, custody, independence, and promotion authorities.

The README and Forge integration documentation now link to the proposal and state clearly that the capability is not yet implemented.

Evidence

• Tests or fixtures cover the change. Not applicable to this documentation-only proposal.
• Specification and documentation agree. No normative specification text is changed; all affected documentation labels the proposal experimental and non-normative.
• make check passes. Not run because this PR was created through the GitHub connector without a local checkout.
• Compatibility and security effects are described.
• Relevant conflicts of interest are disclosed. None known.
• No secrets, private paths, transcripts, caches, or generated build outputs are included.

Diff review confirms three documentation files changed: one new design note plus links in README.md and docs/mncs-forge.md.

RFC

No RFC is required for this PR because it introduces a non-normative design proposal only. Any future schema, conformance, promotion, governance, or normative status changes derived from this work must proceed through the RFC process.

[chatgpt-codex-connector]

You have reached your Codex usage limits for code reviews. You can see your limits in the Codex usage dashboard.
To continue using code reviews, you can upgrade your account or add credits to your account and enable them for code reviews in your settings.