Skip to content

Semantic Notes Vault MCP 0.11.42

Choose a tag to compare

@github-actions github-actions released this 13 Jul 23:49
· 73 commits to main since this release

Semantic Notes Vault MCP 0.11.42

Retrieval signal honesty — follow-up to 0.11.41.

Fixed:

  • Search score labels no longer assert a quality judgement the number cannot support. TF-IDF has no absolute scale, so the fixed bands ('Good' >=1.0, 'Low' >0) were misleading: where most notes share the query term the scores compress, and load-bearing notes were being labelled 'Low'. Scores now render bare and relative to the top hit ('0.36, 25% of top hit'), with an explicit caution not to prune on score.
  • The graph's blind spot is now documented, and it is the mirror of search's. Search misses notes written in different vocabulary; traversal only reaches what someone actually linked. They are complements, not substitutes: scan broadly to catch the unlinked, follow links to catch the differently-worded. The graph description now says so, and points at backlinks for finding what depends on a note.
  • The 'Summary view' footer read as a truncation warning on tools whose job is faithful retrieval. Only metadata is abbreviated; genuine truncation already announces itself. The footer now names what it actually elides.
  • search-traverse is described with its cost (snippets per node, scoreThreshold pruning), not just its brevity.

427 tests passing.

Claude Desktop (.mcpb connector)

If you connect Claude Desktop via the .mcpb bundle, re-import obsidian-mcp.mcpb from this release to update the connector — BRAT updates only the Obsidian plugin, not the Desktop bridge.

Installation via BRAT (Obsidian plugin)

  1. Install the BRAT plugin if you haven't already
  2. Command palette → "BRAT: Add a beta plugin for testing"
  3. Enter: aaronsb/obsidian-mcp-plugin
  4. Enable the plugin in Community Plugins

What's Changed

  • fix(retrieval): stop the score labels lying, and name both blind spots by @aaronsb in #266

Full Changelog: 0.11.41...0.11.42