Skip to content

[deep-report] Embed Hippo Memory store to restore semantic recall quality #28117

@github-actions

Description

@github-actions

The Hippo Memory store has grown to 490 memories (440 episodic, 50 semantic) but only 4 are embedded — a < 1% embedding coverage rate that severely degrades semantic search quality. The store is healthy in all other respects (no stale memories, 0 open conflicts), but semantic queries cannot surface relevant memories because the vector index is nearly empty.

Description: Run hippo embed to embed all 490 memories. Also run hippo audit --fix to prune 7 flagged low-quality entries before embedding to avoid polluting the index. Confirmed in Hippo Memory Insights 2026-04-23.

Expected Impact: Unlocks semantic recall for the entire 490-memory corpus. The store's highest-confidence memory (stale lock files → CI friction, retrieved 3×) is already verified; enabling full semantic search will surface many more contextually relevant lessons during agent runs.

Suggested Agent: Any agent with access to the Hippo MCP server and runner filesystem. Sequential: hippo audit --fixhippo embed.

Estimated Effort: Fast (< 30 min — commands are well-defined, no code changes needed)

Data Source: DeepReport Intelligence Briefing 2026-04-23, Hippo Memory Insights #28007

Generated by DeepReport - Intelligence Gathering Agent · ● 494.2K ·

  • expires on Apr 25, 2026, 3:32 PM UTC

Metadata

Metadata

Assignees

No one assigned

    Type

    No type
    No fields configured for issues without a type.

    Projects

    No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions