feature - migrate the full Oven semantic control plane to Incan authoring

[dannymeijer]

Area

• Tooling (CLI/formatter/test runner)
• Compiler (frontend/backend/codegen)
• Runtime / Core crates (stdlib/core/derive)

Problem statement

One Incan-authored planner proves only a bootstrap. The v0.6 self-hosting-readiness goal requires Oven's complete semantic control plane to be authored in Incan while retaining a deliberately small Rust host kernel for platform and toolchain mechanics.

Proposed solution

Move Oven's semantic project graph, resolver/planner policy, receipt/provenance interpretation, cache/invalidation policy, test/session scheduling, inspection/report construction, and diagnostics policy into Incan modules hosted through the versioned kernel ABI.

Alternatives considered

• Keep all Oven product logic in Rust: leaves Incan authorability unproven.
• Rewrite compiler internals or operating-system mechanics in Incan: exceeds the bounded host-kernel architecture.

Scope / acceptance criteria

• In scope:
  • Migration of every non-kernel Oven responsibility identified by the feature - Incan-authored Oven and toolchain components #870 ownership map.
  • Versioned host-kernel calls, grants, failures, and receipts for each Incan-owned component.
  • Differential behavior, restart, failure-isolation, and no-Cargo normal-operation evidence across supported Oven workflows.
  • An ownership map with no unclassified operational responsibility.
• Out of scope:
  • Compiler frontend self-hosting, general plugins, or arbitrary Rust invocation.
• Done when:
  • Supported Oven workflows are directed by Incan-authored semantic components through the host kernel.
  • Remaining Rust code is a bounded, versioned, justified host capability surface rather than product policy.