QVAC-23953 feat[api]: add injected TurboVec RAG support

[maxim-smotrov]

🎯 What problem does this PR solve?

• RAG cannot use TurboVec without taking a hard runtime dependency on the llama.cpp embedding addon.
• Native candidate search must remain optional while HyperDB stays the durable source of truth.

📝 How does it solve it?

• Adds a provider-based TurboVecAdapter with mutation journaling, checkpoint recovery, locking, and HyperDB-backed scoring.
• Lets the inference embedding plugin expose IdMapIndex through a typed capability that the RAG workspace manager consumes only when QVAC_RAG_TURBOVEC=1.
• Keeps native addon imports inside the plugin boundary.
• This architecture-review draft intentionally excludes generated HyperDB outputs and all release metadata.

🧪 How was it tested?

• @qvac/rag typecheck and lint pass on this feature-only commit.
• RAG unit tests and the inference suite passed during implementation with the generated schema outputs present.
• This draft is not merge-ready because those generated outputs are intentionally omitted.

🔌 API Changes

const adapter = new TurboVecAdapter({
  store,
  indexProvider: {
    create: (options) => new IdMapIndex(options),
    load: (snapshotPath) => IdMapIndex.load(snapshotPath)
  }
})

[github-actions]

License compliance — clean

No new dependency license findings in this PR.

Warn-only (shadow) mode — this check does not block merges yet.

Updated automatically by the canonical license compliance workflow.

NOTICE presence (advisory)

Missing NOTICE (advisory, does not block):

• ./.github/actions/release-merge-guard
• ./docs/website
• ./packages/ggml-coload-smoke
• ./packages/fabric/test/integration
• ./packages/inference-addon-cpp/mobile
• ./packages/sdk/e2e
• ./packages/llm-llamacpp/benchmarks/performance
• ./packages/llm-llamacpp/benchmarks/server
• ./packages/vla-ggml/sim/server
• ./packages/embed-llamacpp/benchmarks/performance
• ./packages/embed-llamacpp/benchmarks/server
• ./packages/asr-ggml/benchmarks/server