#11258: feat(memory): Add Qdrant Vector Database Provider

## Summary This PR adds support for **Qdrant** as a vector memory provider for OpenClaw. ## Motivation Current memory implementations rely on local storage (sqlite-vec), which can be limiting for scalable or cloud-native deployments. Qdrant offers a robust, high-performance solution for managing long-term agent memory in enterprise environments. ## Changes - Added `src/memory/providers/qdrant.ts`: Implements the Qdrant connector with `add` and `search` capabilities. - Added `@qdrant/js-client-rest` dependency. ## Testing - Verified initialization and vector upsert operations locally. <!-- greptile_comment --> <h2>Greptile Overview</h2> <h3>Greptile Summary</h3> This PR introduces a new `QdrantProvider` (Qdrant vector DB client wrapper) intended to support storing and searching embeddings via Qdrant. At the moment, the provider is added as a standalone file but isn’t wired into the existing memory/indexing system (no config/selection/exports), and the implementation hardcodes a 1536-dimension vector size and uses random UUIDs for point IDs—both of which conflict with OpenClaw’s multi-provider embedding dimensions and reindex/dedup semantics. <h3>Confidence Score: 2/5</h3> - This PR is not safe to merge as-is because the added provider isn’t usable and has correctness issues around vector sizing and IDs. - Only a new provider file is added without integration into the memory backend selection, so the feature is effectively absent. Additionally, hardcoding vector dimension to 1536 will fail for non-OpenAI embeddings, and random UUID point IDs break reindex/upsert semantics by creating duplicates. - src/memory/providers/qdrant.ts <!-- greptile_other_comments_section --> _{(4/5) You can add custom instructions or style guidelines for the agent [here](https://app.greptile.com/review/github)!} <!-- /greptile_comment -->