fix(db): normalize agent name in memory vector search

[mesutoezdil]

ListAgentMemories and DeleteAgentMemory match both the hyphen and underscore forms of the agent name. SearchAgentMemory only matched the raw form, so a search with one form silently missed memories stored under the other. The same normalization is now applied in the search query. Covered by a new test.

[Copilot]

Pull request overview

This PR fixes inconsistent agent-name normalization in the Postgres vector-search query so SearchAgentMemory returns memories regardless of whether they were stored under a hyphenated or underscore-normalized agent name. This aligns search behavior with the existing normalization behavior in list/delete paths and adds a regression test.

Changes:

• Update SearchAgentMemory SQL to match either of two agent-name variants (raw + normalized) while keeping the user scoping intact.
• Update the Postgres DB client to pass both the raw agent name and its hyphen→underscore normalized variant into the query.
• Add a new test that stores a memory under the underscore form and verifies searching with the hyphenated form returns it.

Reviewed changes

Copilot reviewed 3 out of 4 changed files in this pull request and generated no comments.

File	Description
go/core/internal/database/queries/memory.sql	Expands SearchAgentMemory filtering to match either raw or normalized agent name, adjusts parameter ordering.
go/core/internal/database/gen/memory.sql.go	Regenerates sqlc output to reflect the updated query signature (additional agent-name parameter).
go/core/internal/database/client_postgres.go	Passes both raw and normalized agent name into SearchAgentMemory.
go/core/internal/database/client_test.go	Adds regression test for normalized-name search and includes minor comment punctuation tweaks.

Files not reviewed (1)

• go/core/internal/database/gen/memory.sql.go: Generated file

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.