Audit SDK public API for no_std cognitive runtime constraints

[AxonOS-BCI]

Context

AxonOS SDK is intended to expose a minimal, deterministic, low-latency API surface for cognitive runtime components.

This issue is a first review task for checking whether the current SDK public interface is aligned with a no_std, systems-level, low-allocation runtime model.

Scope

Review the current public API surface and identify anything that may conflict with the AxonOS SDK direction:

• unnecessary std assumptions;
• avoidable heap allocation requirements;
• overly generic module or type names;
• unclear module boundaries;
• APIs that look too broad for a minimal cognitive runtime;
• naming that does not clearly belong to AxonOS;
• dependencies that may be inappropriate for embedded / bare-metal / low-latency environments.

Non-goals

Do not rewrite the architecture.

Do not introduce new dependencies.

Do not change public APIs without discussion.

Do not add large abstractions.

Do not convert this into a general AI-agent framework review.

Expected output

Submit one of the following:

1. A short Markdown report as an issue comment; or
2. A small pull request via fork if the proposed changes are minimal and obvious.

The review should include:

• files reviewed;
• concrete concerns;
• suggested minimal changes;
• any uncertainty or open questions.

Acceptance criteria

This issue is complete when the review provides:

• a clear list of API-surface concerns;
• specific references to modules, files, functions, or types;
• minimal recommendations aligned with no_std, low-latency, privacy-preserving AxonOS SDK design.