Add entity-label filtering and developer docs

[cpdata]

Summary

• extend the GraphDriver contract and every driver implementation to honor namespace and entity-label filters, wiring the behaviour through MemoryManager, the MeshMind client, and REST/gRPC search/list endpoints
• build a developer-facing wiki under docs/ plus ENVIRONMENT_NEEDS.md and RESUME_NOTES.md, and update contributor instructions so documentation stays synchronized with code changes
• expand regression coverage and examples so SQLite/fake drivers, service stubs, and the CLI demo all exercise the new filtering semantics

Testing

• pytest

---

https://chatgpt.com/codex/tasks/task_b_68ee1c9ee92c83218cbbebce8b0667b8

[chatgpt-codex-connector]

💡 Codex Review

Here are some automated review suggestions for this pull request.

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

• Open a pull request for review
• Mark a draft as ready
• Comment "@codex review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

Codex can also answer questions or update the PR. Try commenting

• @codex fix this CI failure
• @codex address that feedback

[chatgpt-codex-connector]

Use ranked results for reranker subset

The hybrid search path builds baseline from hybrid_search, but the reranker is invoked with mems[:cfg.rerank_k]. mems is the unranked, filtered input order, so the LLM reranker operates on arbitrary memories instead of the top-scoring candidates from the hybrid search. In practice the reranker can ignore the best matches and reorder unrelated items, degrading result quality. The reranker should consume baseline[:cfg.rerank_k] (the ranked list) before merging back into ordered.

Useful? React with 👍 / 👎.