# Local Vector Search MCP (nuget · DimonSmart.LocalVectorSearchMcp)

Local Markdown search, navigation and editing via MCP, with optional vector embeddings.

- Trust score: 67/100 (medium)
- Change this week: +5
- Registry status: active
- Liveness: live
- Owner verified: no
- Last scored: 2026-10-24

## Components

- nuget · `DimonSmart.LocalVectorSearchMcp`: 67/100 (this document), [markdown](https://verifymcp.io/servers/dimonsmart-local-vector-search-mcp/dimonsmart-localvectorsearchmcp.md), [page](https://verifymcp.io/servers/dimonsmart-local-vector-search-mcp/dimonsmart-localvectorsearchmcp)

## Channel facts

- Registry: `nuget`
- Package: `DimonSmart.LocalVectorSearchMcp`
- Version: `0.1.3`
- Transport: `stdio`

## Trust breakdown

How this component scores in each security and reliability category. Every signal is checked automatically from public evidence about the published package, including repeated runs of it in an isolated sandbox, and we only credit what we can confirm. Scores are 0–100 per category. Scoring method: https://verifymcp.io/docs/scoring (what has changed: https://verifymcp.io/docs/scoring/changelog)

Scored 2026-10-24.

- **Supply Chain Security**: 88/100
  - No malware found by supply-chain analysis.
  - No known CVEs affecting this package version or its production dependencies.
  - Install-script risk not yet assessed.
  - No production dependencies, so there is no dependency health to assess.
- **Provenance & Transparency**: 45/100
  - Source repository is publicly reachable at the declared URL.
  - Build provenance not yet verified.
  - Clear OSI-approved license (MIT).
  - Actively maintained (last published 12 days ago).
  - Disclosure check failed: no security disclosure policy was found in the source repository.
- **Schema Quality & AI Usability**: 67/100
  - AI-judged instruction clarity (good).
  - Tool/resource definitions use about 1885 tokens (~104/item across 18 items; 18 tools + 0 resources), lean.
  - Usage-examples check failed: none of the tools include examples.
- **Stability & Change Management**: 33/100
  - Stability observed for 10 of 30 days with no destabilising changes; credit accrues until the full window elapses.
- **Tool Coverage**: 79/100
  - 100% of tools have a non-trivial description (not blank, and not just the tool's name).
  - 27% of tool parameters carry a description.
  - Structured output schemas are declared (89% of tools); any adoption earns full credit.
- **Tool Safety**: 75/100
  - No prompt-injection markers were found in the server instructions, tool names or descriptions we captured.
  - 0 of 2 tool(s) whose name or description implies an irreversible operation declare an MCP destructiveHint annotation; "kb_delete_image" implies "delete" and declares no destructiveHint at all, which the MCP spec reads as destructive by default.
  - An AI judge read all 18 captured unit(s) of tool text and found none that tries to manipulate the model reading it.
- **Capabilities**: 100/100
  - Implements a current MCP spec version (2026-07-28).

## Install

### How do I install the Local Vector Search MCP server?

Local Vector Search MCP runs locally as a NuGet package, launched with dnx DimonSmart.LocalVectorSearchMcp@0.1.3 --yes. Ready-made configuration for Claude, Cursor, VS Code, Codex and 3 more is on this page, copied from each client's own documentation.

### Claude

```bash
claude mcp add dimonsmart-local-vector-search-mcp -- dnx DimonSmart.LocalVectorSearchMcp@0.1.3 --yes
```

### Cursor

```json
{
  "mcpServers": {
    "dimonsmart-local-vector-search-mcp": {
      "command": "dnx",
      "args": [
        "DimonSmart.LocalVectorSearchMcp@0.1.3",
        "--yes"
      ]
    }
  }
}
```

### VS Code

```json
{
  "servers": {
    "dimonsmart-local-vector-search-mcp": {
      "command": "dnx",
      "args": [
        "DimonSmart.LocalVectorSearchMcp@0.1.3",
        "--yes"
      ]
    }
  }
}
```

### Codex

```bash
codex mcp add dimonsmart-local-vector-search-mcp -- dnx DimonSmart.LocalVectorSearchMcp@0.1.3 --yes
```

### opencode

```json
{
  "$schema": "https://opencode.ai/config.json",
  "mcp": {
    "dimonsmart-local-vector-search-mcp": {
      "type": "local",
      "command": [
        "dnx",
        "DimonSmart.LocalVectorSearchMcp@0.1.3",
        "--yes"
      ],
      "enabled": true
    }
  }
}
```

### Hermes

```yaml
mcp_servers:
  dimonsmart-local-vector-search-mcp:
    command: "dnx"
    args: ["DimonSmart.LocalVectorSearchMcp@0.1.3", "--yes"]
```

### Netclaw

```json
{
  "McpServers": {
    "dimonsmart-local-vector-search-mcp": {
      "Transport": "stdio",
      "Command": "dnx",
      "Arguments": [
        "DimonSmart.LocalVectorSearchMcp@0.1.3",
        "--yes"
      ]
    }
  }
}
```

### Other

```json
{
  "mcpServers": {
    "dimonsmart-local-vector-search-mcp": {
      "command": "dnx",
      "args": [
        "DimonSmart.LocalVectorSearchMcp@0.1.3",
        "--yes"
      ]
    }
  }
}
```

## Changelog

Every change recorded for this component, newest first. Days that predate change tracking, or that we cannot explain, say so: "we were watching and nothing happened" and "we were not watching" are different claims.

### 2026-10-24 (score 67, +1)

No change was recorded against any check on this day. Stability & Change Management went from 30 to 33. That category is still filling its 30-day observation window: 9 days of observed history at the previous scan, 10 at this one. The score rises as the window fills, whether or not the server changes.

### 2026-10-22 (score 66, +4)

- [functional improvement] Stability: unverified → 0.27

### 2026-10-14 (score 62, +27)

- [security improvement] Injection markers: unverified → pass
- [security] First check of Judged manipulation: pass
- [security] Stability: Stability not yet verified: not enough scan history yet (needs a 30-day window).
- [functional improvement] Tool coverage: unverified → 100
- [functional improvement] MCP protocol: unverified → pass
- [functional] First check of Schema quality: fail
- [functional] First check of Destructive annotations: 0
- [functional] First check of Schema quality: pass
- [functional] First check of Schema quality: good
- [functional] First check of Tool coverage: 27
- [functional] First check of Tool coverage: 89

### 2026-10-12 (score 35)

First indexed and scored.

## MCP tools (18)

### `kb_list_files` (~51 tokens)

Lists Markdown and other permitted asset files under the configured workspace root. pathPrefix limits the search area to one directory or one exact file; use includeGlob to match filename patterns.

Input parameters:

- `request` (object, required)

Output parameters:

- `files` (array)

### `kb_patch` (~307 tokens)

Atomically edits Markdown using hashed semantic anchors, commits the source file, and schedules index reconciliation. Public concrete pointers use logical~selfHash~subtreeHash with 16-character lowercase hashes. For list_item, replace_element (and deprecated replace) modifies only parent-owned content before nested items, preserving children; it checks selfHash. replace_subtree and delete explicitly modify the entire list subtree and require both hashes; insert_before/insert_after insert one sibling and check only selfHash. Block quotes are opaque qN elements with no separately addressable children. Other self-scoped operations validate selfHash. replace is a deprecated alias of replace_element. replace_subtree applies only to list_item, including its descendants. delete removes exactly one Markdown element; for a heading, the section body is preserved. replace_section replaces a heading and its complete owned section subtree; delete_section removes a heading and its complete owned section subtree. Section operations run through the next heading with level <= the target level or end of document, require a canonical v2 heading pointer, and validate subtreeHash so stale body or raw-content edits cannot be overwritten or deleted. Relocation uses only exact element kind plus selfHash. Legacy logical~selfHash input remains accepted for Self operations but is rejected for replace_section and delete_section with a reread instruction. The document pointer remains unhashed and supports insert_before and insert_after at document boundaries. IndexSynchronized reports whether derived-index synchronization has already been confirmed.

Input parameters:

- `request` (object, required)

Output parameters:

- `indexError` (string|null)
- `indexSynchronized` (boolean)
- `path` (string)
- `previousPath` (string|null)
- `sourceHash` (string|null)

### `kb_create_pdf_book` (~121 tokens)

Creates a formatted PDF book from one Markdown document. Supports headings, formatting, code, lists, quotes, links, GFM tables and local images. Requires knowledgeBase.allowWrites=true; output is root-relative, ends in .pdf and replaces an existing PDF atomically. Local images outside the workspace or unavailable images become visible placeholders and rendering warnings. Optional profile screen, standard (default), or print controls image DPI and JPEG quality; print is not PDF/X. Markdown, image assets and the search index are not modified.

Input parameters:

- `request` (object, required)

Output parameters:

- `bytes` (integer)
- `path` (string)
- `warnings` (array)

### `kb_reindex` (~51 tokens)

Starts reindexing of the configured Markdown root in the background. Only one reindex can run at a time. Use kb_status to monitor progress and obtain the final result.

Input parameters:

- `request` (object, required)

Output parameters:

- `current` (object): Active operation, including the previously running operation when started is false. Check kb_status for completion.
- `started` (boolean): True if a new operation was queued; false if an operation was already running. This does not mean indexing has completed.

### `kb_download_file` (~87 tokens)

Returns one permitted root-relative workspace file as an embedded binary resource, followed by JSON metadata. Accepts any regular file up to 25 MiB. Git metadata, internal index files, absolute/traversal paths, and paths through reparse points are not permitted. Does not require writes and does not modify the workspace or index.

Input parameters:

- `path` (string, required): Root-relative workspace file path.

### `kb_list_images` (~69 tokens)

Lists supported images recursively under knowledgeBase.root, excluding Git and index files; cursor pagination defaults to 50.

Input parameters:

- `cursor` (string|null): Opaque cursor returned by the previous page.
- `pageSize` (integer|null): Page size from 1 to 200. Defaults to 50.

Output parameters:

- `images` (array)
- `nextCursor` (string|null)

### `kb_search` (~82 tokens)

Searches the derived Markdown index using lexical, semantic or hybrid search. Results belong to an indexed revision identified by indexedSourceHash and may temporarily lag behind the current source. Search pointers refer to the first element of a chunk, which may differ from the matching list item. Hybrid search falls back to lexical search when the embedding provider is unavailable.

Input parameters:

- `request` (object, required)

Output parameters:

- `results` (array)
- `warning` (string|null): Potential search fallback or limitation; null when no warning applies.

### `kb_delete_image` (~47 tokens)

Deletes one supported binary image under knowledgeBase.root without checking or changing Markdown references, index or parent directories. Requires allowWrites=true.

Input parameters:

- `path` (string, required): Root-relative image file path.

Output parameters:

- `deleted` (boolean)
- `path` (string)

### `kb_create` (~50 tokens)

Creates a new UTF-8 Markdown file, commits it as the source of truth, and schedules index synchronization. IndexSynchronized reports whether derived-index synchronization has already been confirmed.

Input parameters:

- `request` (object, required)

Output parameters:

- `indexError` (string|null)
- `indexSynchronized` (boolean)
- `path` (string)
- `previousPath` (string|null)
- `sourceHash` (string|null)

### `kb_load_image` (~92 tokens)

Returns an MCP image content block and a separate text content block containing JSON metadata (path, mimeType, bytes and sha256 of the original file). GIF images produce a PNG preview of the first frame without animation; JSON metadata still describes the original GIF. The complete multimodal result is not represented by a JSON outputSchema.

Input parameters:

- `path` (string, required): Root-relative image path, for example chapters/diagram.png.

### `kb_move` (~52 tokens)

Moves a Markdown file when its source revision matches, commits the filesystem move, and schedules index reconciliation for the affected paths. IndexSynchronized reports whether derived-index synchronization has already been confirmed.

Input parameters:

- `request` (object, required)

Output parameters:

- `indexError` (string|null)
- `indexSynchronized` (boolean)
- `path` (string)
- `previousPath` (string|null)
- `sourceHash` (string|null)

### `kb_outline` (~25 tokens)

Returns a deterministic heading outline for one Markdown file.

Input parameters:

- `request` (object, required)

Output parameters:

- `headings` (array)
- `path` (string)
- `sourceHash` (string)

### `kb_edit_table` (~345 tokens)

Edits one GFM pipe table without generating whole-table Markdown. First call kb_read on the table and pass its full hashed tN pointer. update_cells requires nonempty updates; every update requires value, one row selector (rowIndex or where), and one column selector (column or columnIndex). insert_row requires values or valuesByColumn; delete_row requires a row selector. insert_column requires name and accepts defaultValue, alignment, and an optional before-column selector. delete_column requires a column selector; rename_column also requires newName; set_alignment requires alignment (none, left, center, or right). Data rows have zero-based rowIndex; header is excluded. Prefer unique where {column,equals} selectors and column names; use numeric indexes if names or values are ambiguous. Irrelevant or conflicting action fields are rejected. Use the returned pointer for subsequent edits; on CONFLICT call kb_read again. Use kb_patch to delete the entire table.

Input parameters:

- `action` (string, required)
- `alignment` (string|null)
- `beforeColumn` (string|null)
- `beforeColumnIndex` (integer|null)
- `beforeRowIndex` (integer|null)
- `column` (string|null)
- `columnIndex` (integer|null)
- `defaultValue` (string|null)
- `name` (string|null)
- `newName` (string|null)
- `path` (string, required)
- `pointer` (string, required)
- `rowIndex` (integer|null)
- `updates` (array|null)
- `values` (array|null)
- `valuesByColumn` (object|null)
- `where` (object|null)

Output parameters:

- `indexError` (string|null)
- `indexSynchronized` (boolean)
- `path` (string)
- `pointer` (string)
- `sourceHash` (string)

### `kb_move_image` (~80 tokens)

Moves a supported image only within the configured knowledgeBase.root. Paths outside the configured root are never allowed. Never updates Markdown references or the index, and never overwrites an existing target. Repair links separately with kb_patch.

Input parameters:

- `request` (object, required): Binary image move using root-relative sourcePath and targetPath; optional expectedSha256 checks for source changes.

Output parameters:

- `path` (string)
- `previousPath` (string)
- `sha256` (string)

### `kb_delete` (~51 tokens)

Deletes a Markdown file when its source revision matches, commits the source deletion, and schedules removal from the derived index. IndexSynchronized reports whether derived-index synchronization has already been confirmed.

Input parameters:

- `request` (object, required)

Output parameters:

- `indexError` (string|null)
- `indexSynchronized` (boolean)
- `path` (string)
- `previousPath` (string|null)
- `sourceHash` (string|null)

### `kb_save_image` (~190 tokens)

Saves a PNG, JPEG, WebP, or GIF supplied through the client-supported OpenAI file-parameter mechanism inside knowledgeBase.root (by default images/). Upload or select the image in the client and pass its resulting file reference; arbitrary text, raw Base64, and data URIs are not supported. Requires knowledgeBase.allowWrites=true and never overwrites an existing image.

Input parameters:

- `altText` (string|null): Optional alt text used to build the returned Markdown image reference.
- `file` (object, required): Actual image file reference supplied through the client's OpenAI file-parameter mechanism. Do not provide arbitrary text, raw Base64, or a data URI.
- `fileName` (string|null): Optional explicit portable basename for the saved image. Directories are not allowed.
- `targetPath` (string|null): Optional exact root-relative destination including the image file name. Cannot be combined with fileName.

Output parameters:

- `bytes` (integer)
- `markdown` (string)
- `mimeType` (string)
- `path` (string)
- `sha256` (string)

### `kb_status` (~33 tokens)

Returns local vector search index status, including project counts, pending synchronization, active or last reindex operation, and manifest compatibility.

Output parameters:

- `chunkerVersion` (string)
- `compatibility` (object|null)
- `databasePath` (string)
- `embeddingDimensions` (integer|null)
- `embeddingModel` (string)
- `embeddingProvider` (string)
- `embeddingTextBuilderVersion` (string)
- `indexMode` (string)
- `indexing` (object|null)
- `project` (object)
- `schemaVersion` (string)
- `synchronization` (object|null)

### `kb_read` (~152 tokens)

Reads the current Markdown source starting at a semantic pointer or hashed semantic anchor. Reading is independent of embeddings and background indexing. Public concrete pointers are returned as logical~selfHash~subtreeHash; legacy logical~selfHash input remains valid for navigation. List items are liN addresses, nested within parent liN; BlockQuote is one opaque qN element. GFM tables are standalone tN elements; address a table to receive structured columns and data rows. Table cells are normalized display text (for example, inline Markdown formatting is removed); use the returned markdown slice when original cell markup is needed. Omit pointer or use "document" to read from the beginning of the document.

Input parameters:

- `request` (object, required)

Output parameters:

- `elements` (array)
- `markdown` (string)
- `nextPointer` (string|null)
- `path` (string)
- `pointer` (string)
- `sourceHash` (string)
- `table` (object|null)

## Diagnostics

Captured diagnostic sections: Provenance, Dependencies. The full working is on the page: https://verifymcp.io/servers/dimonsmart-local-vector-search-mcp/dimonsmart-localvectorsearchmcp#diagnostics

## Score history

- 2026-10-24: 67
- 2026-10-23: 66
- 2026-10-22: 66
- 2026-10-21: 62
- 2026-10-20: 62
- 2026-10-19: 62
- 2026-10-18: 62
- 2026-10-17: 62
- 2026-10-16: 62
- 2026-10-15: 62
- 2026-10-14: 62
- 2026-10-13: 35
- 2026-10-12: 35

## Common questions

### What is the Local Vector Search MCP server?

Local Vector Search MCP is listed in the public MCP registry as io.github.DimonSmart/local-vector-search-mcp. Local Markdown search, navigation and editing via MCP, with optional vector embeddings. This page covers its NuGet package (DimonSmart.LocalVectorSearchMcp).

### Is the Local Vector Search MCP server safe to use?

Local Vector Search MCP scores 67 out of 100 on VerifyMCP. We found no known CVEs affecting it as of 24 October 2026. That is a record of what we were able to check automatically, not an endorsement. The category breakdown on this page shows every signal behind the number, including the ones we could not confirm.

### What tools does the Local Vector Search MCP server expose?

Local Vector Search MCP exposes 18 tools: kb_list_files, kb_patch, kb_create_pdf_book, kb_reindex, kb_download_file, and 13 more. Their descriptions and schemas cost roughly 1,885 tokens of context every time the server is loaded.

### Is the Local Vector Search MCP server still maintained?

Local Vector Search MCP is still listed as active in the MCP registry. We last reached this channel on 24 October 2026. Those dates come from our own scans of the registry and the channel itself, not from anything the publisher announced.

### What licence is the Local Vector Search MCP server under?

Local Vector Search MCP declares the MIT licence, which is OSI-approved. That covers the source only, and says nothing about the cost of any service it calls.

## Links

- NuGet package: https://www.nuget.org/packages/DimonSmart.LocalVectorSearchMcp
- Socket report: https://socket.dev/nuget/package/DimonSmart.LocalVectorSearchMcp
- Repository: https://github.com/DimonSmart/LocalVectorSearchMcp
- Changelog RSS feed: https://verifymcp.io/servers/dimonsmart-local-vector-search-mcp/dimonsmart-localvectorsearchmcp.xml
- Changelog JSON feed: https://verifymcp.io/servers/dimonsmart-local-vector-search-mcp/dimonsmart-localvectorsearchmcp.json
- HTML version of this page: https://verifymcp.io/servers/dimonsmart-local-vector-search-mcp/dimonsmart-localvectorsearchmcp