# io.github.Newmannn91/memdrive-mcp (npm · memdrive-mcp)

Local-first MCP memory: Markdown notes and images with full-text search for Claude and Cursor.

- Trust score: 70/100 (medium)
- Change this week: +4
- Registry status: active
- Liveness: live
- Owner verified: no
- Last scored: 2026-11-21

## Components

- npm · `memdrive-mcp`: 70/100 (this document), [markdown](https://verifymcp.io/servers/newmannn91-memdrive-mcp/memdrive-mcp.md), [page](https://verifymcp.io/servers/newmannn91-memdrive-mcp/memdrive-mcp)

## Channel facts

- Registry: `npm`
- Package: `memdrive-mcp`
- Version: `0.9.2`
- 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 2026-11-21.

- **Supply Chain Security**: 90/100
  - No malware found by supply-chain analysis.
  - CVE check failed: a known high-severity CVE affects zod 3.25.76, reached via jimp > @jimp/plugin-blit > zod. A fixed version is available.
  - No install/post-install scripts declared.
  - 50 of 154 dependencies flagged as unhealthy.
- **Provenance & Transparency**: 45/100
  - Source repository is publicly reachable at the declared URL.
  - Provenance check failed: no build-provenance attestation is published.
  - Clear OSI-approved license (MIT).
  - Actively maintained (last published 4 days ago).
  - Disclosure check failed: no security disclosure policy was found in the source repository.
- **Schema Quality & AI Usability**: 64/100
  - AI-judged instruction clarity (excellent).
  - Context-footprint check failed: tool/resource definitions use about 4370 tokens (~397/item across 11 items; 11 tools + 0 resources), over budget; trim descriptions and params.
  - Usage-examples check failed: none of the tools include examples.
- **Stability & Change Management**: 40/100
  - Stability observed for 12 of 30 days with no destabilising changes; credit accrues until the full window elapses.
- **Tool Coverage**: 100/100
  - 100% of tools have a non-trivial description (not blank, and not just the tool's name).
  - 100% of tool parameters carry a description.
- **Tool Safety**: 75/100
  - No prompt-injection markers were found in the server instructions, tool names or descriptions we captured.
  - 0 of 1 tool(s) whose name or description implies an irreversible operation declare an MCP destructiveHint annotation; "apply_plan" implies "execute" and declares no destructiveHint at all, which the MCP spec reads as destructive by default.
  - An AI judge read all 11 captured unit(s) of tool text and found none that tries to manipulate the model reading it.
- **Capabilities**: 100/100
  - Implements a supported MCP spec version (2025-11-25); the latest is 2026-07-28.

## Install

### How do I install the io.github.Newmannn91/memdrive-mcp server?

io.github.Newmannn91/memdrive-mcp runs locally as an npm package, launched with npx -y memdrive-mcp. Ready-made configuration for Claude, Cursor, VS Code, Codex and 5 more is on this page, copied from each client's own documentation.

### Claude

```bash
claude mcp add newmannn91-memdrive-mcp -- npx -y memdrive-mcp
```

### Cursor

```json
{
  "mcpServers": {
    "newmannn91-memdrive-mcp": {
      "command": "npx",
      "args": [
        "-y",
        "memdrive-mcp"
      ]
    }
  }
}
```

### VS Code

```json
{
  "servers": {
    "newmannn91-memdrive-mcp": {
      "command": "npx",
      "args": [
        "-y",
        "memdrive-mcp"
      ]
    }
  }
}
```

### Codex

```bash
codex mcp add newmannn91-memdrive-mcp -- npx -y memdrive-mcp
```

### opencode

```json
{
  "$schema": "https://opencode.ai/config.json",
  "mcp": {
    "newmannn91-memdrive-mcp": {
      "type": "local",
      "command": [
        "npx",
        "-y",
        "memdrive-mcp"
      ],
      "enabled": true
    }
  }
}
```

### OpenClaw

```bash
openclaw mcp add newmannn91-memdrive-mcp --command npx --arg -y --arg memdrive-mcp
```

### Hermes

```yaml
mcp_servers:
  newmannn91-memdrive-mcp:
    command: "npx"
    args: ["-y", "memdrive-mcp"]
```

### Netclaw

```json
{
  "McpServers": {
    "newmannn91-memdrive-mcp": {
      "Transport": "stdio",
      "Command": "npx",
      "Arguments": [
        "-y",
        "memdrive-mcp"
      ]
    }
  }
}
```

### Vellum

```bash
assistant mcp add newmannn91-memdrive-mcp -t stdio -c npx -a -y memdrive-mcp
```

### Other

```json
{
  "mcpServers": {
    "newmannn91-memdrive-mcp": {
      "command": "npx",
      "args": [
        "-y",
        "memdrive-mcp"
      ]
    }
  }
}
```

## 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.

### 2026-11-21 (score 70, +1)

No change was recorded against any check on this day. Stability & Change Management went from 37 to 40. That category is still filling its 30-day observation window: 11 days of observed history at the previous scan, 12 at this one. The score rises as the window fills, whether or not the server changes. Other categories moved too: Supply Chain Security rose 1.

### 2026-11-20 (score 69, +1)

No change was recorded against any check on this day. Stability & Change Management went from 33 to 37. That category is still filling its 30-day observation window: 10 days of observed history at the previous scan, 11 at this one. The score rises as the window fills, whether or not the server changes.

### 2026-11-17 (score 68, +9)

- [security regression] Known CVEs: unverified → fail
- [functional improvement] Dependency health: unverified → 0.86

### 2026-11-16 (score 59, −8)

- [security regression] Stability: 0.20 → unverified
- [security regression] Tool safety: pass → unverified
- [security regression] Known CVEs: fail → unverified
- [functional regression] Capabilities: pass → unverified
- [functional regression] Dependency health: 0.86 → unverified
- [functional regression] Tool coverage: 100 → unverified
- [functional] First check of Schema quality: unverified
- [functional] Package version: 0.9.0 → 0.9.2

### 2026-11-15 (score 67, +1)

No change was recorded against any check on this day. Stability & Change Management went from 17 to 20. That category is still filling its 30-day observation window: 5 days of observed history at the previous scan, 6 at this one. The score rises as the window fills, whether or not the server changes.

### 2026-11-14 (score 66, −1)

- [functional improvement] Schema quality: 513 → 394
- [functional] Destructive annotations: pass → 0
- [functional] Package version: 0.8.4 → 0.9.0

### 2026-11-12 (score 67, +16)

- [security improvement] Malware scan: unverified → pass

### 2026-11-11 (score 51, −15)

- [security regression] Malware scan: pass → unverified
- [functional] Package version: 0.8.2 → 0.8.4
- [functional] Package version: 0.8.2 → 0.8.3

## MCP tools (11)

### `list_directory` (~120 tokens)

List the immediate children (subfolders and files) of a directory inside the MemDrive memory library.
Returns one line per entry with type, size, and modification time. Directories sort first.
Use this to explore what is already stored before reading or saving.
Paths are relative to the library root (MEMDRIVE_ROOT). Omit path or pass "" for the root. Absolute paths are accepted only when they stay inside the root.

Input parameters:

- `path` (string): Directory to list, relative to the library root. Omit or pass "" for the root.

### `search_files` (~749 tokens)

Search the memory library by keyword, ranked by relevance (full-text index + BM25; Chinese queries are segmented, so 1–2 character terms match).
Returns one hit per matching file (its best-matching chunk) with the file path, section heading, relevance score, and a snippet with the matched term highlighted as **term**. Filename matches are included too: a substring of a filename finds the file, including non-text files such as images. Results are paginated: limit defaults to 20, pass offset for the next page.
Filters: file_types (e.g. [".md"]), tags (a file must carry at least one), date_from/date_to (file date: front matter > filename > modification time; a date-only upper bound covers the whole day), modified_after/modified_before (modification time). exact_phrase requires the phrase to appear verbatim.
diagnostics_only returns a read-only library report (empty files, duplicate content, broken image references, unreferenced attachments older than 24h) and never modifies anything; keyword is not needed for it. rebuild_index forces a full index rebuild first (use after a bulk import or a cache clear).
On the first search after startup the index may still be building in the background; the result then says so and falls back to a direct scan (keyword, path, and file types only — tags, dates, and exact_phrase are not applied there, and the result says so).
Paths are relative to the library root (MEMDRIVE_ROOT). Omit path or pass "" to search the whole library.
Files under the top-level archive/ folder are excluded from search_files — whole-library search and an explicit path: "archive" search both return nothing for them — while list_directory still shows archive/ so archived files remain inspectable.

Input parameters:

- `date_from` (string): Earliest file date (ISO date, e.g. "2026-01-01"). File date is front matter date, then filename date, then modification time. A date-only value means that calendar day in the server's local timezone.
- `date_to` (string): Latest file date (ISO date, e.g. "2026-12-31"). A date-only value covers that whole calendar day in the server's local timezone.
- `diagnostics_only` (boolean): Return the read-only library diagnostics report instead of searching (default false).
- `exact_phrase` (string): Phrase that must appear verbatim in a matching chunk (in addition to the keyword).
- `file_types` (array): File extensions to include, e.g. [".md", ".txt"]. Omit for all types.
- `keyword` (string): Keyword to search for (case-insensitive; Chinese terms are segmented). Required unless diagnostics_only is true.
- `limit` (integer): Maximum files to return (default 20, max 200).
- `modified_after` (string): Earliest modification time (ISO date). A date-only value means the start of that calendar day in the server's local timezone.
- `modified_before` (string): Latest modification time (ISO date). A date-only value covers that whole calendar day in the server's local timezone.
- `offset` (integer): Files to skip for pagination (default 0).
- `path` (string): Directory to search under, relative to the library root. Omit or pass "" for the whole library.
- `rebuild_index` (boolean): Force a full index rebuild before searching (default false).
- `tags` (array): A file must carry at least one of these tags (YAML front matter `tags:` or body `#tag`). Omit for all.

### `save_file` (~935 tokens)

CLIPBOARD FIRST — READ THIS BEFORE ANYTHING ELSE: When the user says to save a screenshot or an image from the clipboard — for example "把截图存下来", "把剪贴板里的图存下来", "save the screenshot", or any equivalent wording — pass images: ["clipboard"]. That is all it takes: the server reads the clipboard directly on the user's own machine. You do NOT need to see the clipboard or the image. NEVER ask the user to paste, upload, or send the image into the chat. NEVER say you cannot read the clipboard, that you cannot see images, or that this memory library can only store text. If the clipboard holds no image, the server reports that in the result and the text is still saved — tell the user to take the screenshot again and retry; do not invent a clipboard failure.
Interaction rule — read before calling:
Before calling this tool, show the user the EXACT destination path and
filename you intend to write (for example: "notes/2026-10-28.md") and
get their explicit confirmation in the chat. Never call save_file with
a path or filename the user has not confirmed. If the user has not
confirmed a location yet, ask them first instead of calling.
There is no popup or confirmation dialog: confirmation happens here,
in the conversation, through you.
Save text (usually Markdown) into the MemDrive memory library.
Modes: "new" (default) fails when the file already exists; "overwrite" replaces it; "append" adds to the end, creating the file when missing. Parent directories are created automatically.
Tags: two syntaxes are supported — YAML front matter `tags:` (inline `[a, b]` or block `- a`) and body `#tag`. Proactively suggest adding tags ONLY in these three cases: (1) the user mentions a category or topic while saving (e.g. saving a work plan); (2) the note clearly belongs to a tag already used in the library; (3) the user saves several notes on the same topic in a row (ask once, e.g. whether to tag all of them X). NEVER ask about tags for quick screenshot/clipboard saves, when the user only said…

Input parameters:

- `content` (string, required): Text content to save (usually Markdown).
- `images` (array): Images to archive with the document — CLIPBOARD FIRST: the literal string "clipboard" captures the latest screenshot/image from the OS clipboard on the machine running the server. Use it whenever the…
- `mode` (string): Write mode: "new" (default, fails if the file exists), "overwrite", or "append".
- `path` (string, required): Destination file path, relative to the library root (for example "notes/2026-10-28.md"). Must be confirmed with the user before calling.

### `edit_tags` (~420 tokens)

Edit the tags on an existing note: add one tag, remove one tag, or rename one tag. One operation per call.
Tags live in two places: YAML front matter `tags:` (inline `[a, b]` or block `- a`) and body `#tag`s. The location parameter picks which one to edit (default front_matter).
Only the tag lines change: the rest of the file is left byte-for-byte identical (no reflow, no other front-matter field touched). Adding a tag to a note with no front matter creates a minimal `---\ntags: [tag]\n---` block at the top; adding a body tag appends `#tag` on its own line at the end of the file.
No-ops are not errors: adding a tag that is already there, removing one that is not there, and renaming to a name that already exists all succeed with a "No change" result. Tag names are case-insensitive.
The search index is updated at once, so the note is searchable by its new tags immediately.
Paths are relative to the library root (MEMDRIVE_ROOT). Absolute paths are accepted only when they stay inside the root. The root is a hard sandbox: "../" escapes and symlinks pointing outside are rejected.

Input parameters:

- `location` (string): Which tag syntax to edit: "front_matter" (default, YAML `tags:`) or "body" (`#tag`s in the text).
- `new_tag` (string): Required for "rename": the new tag name. Ignored for "add" and "remove".
- `operation` (string, required): One operation per call: "add" adds tag, "remove" removes tag, "rename" renames tag to new_tag.
- `path` (string, required): Note to edit, relative to the library root (for example "notes/2026-10-28.md").
- `tag` (string, required): The tag to add, remove, or rename (without the leading "#").

### `archive_file` (~213 tokens)

Move a file into the top-level `archive/` folder, preserving its relative path (`notes/foo.md` becomes `archive/notes/foo.md`).
Archiving is recoverable and is the default way to retire old or superseded files instead of deleting them. Use `restore_file` to move an archived file back.
Files in `archive/` are excluded from `search_files`, while `list_directory` still shows `archive/` so archived files remain inspectable.
Refuses to overwrite if the destination under `archive/` already exists, and rejects directories, missing files, `archive/` itself, or files already inside `archive/`.
Paths are relative to the library root (MEMDRIVE_ROOT). Absolute paths are accepted only when they stay inside the root. The root is a hard sandbox: "../" escapes and symlinks pointing outside are rejected.

Input parameters:

- `path` (string, required): File to archive, relative to the library root (for example "notes/2026-10-28.md").

### `restore_file` (~169 tokens)

Restore an archived file from the top-level `archive/` folder back to its original path (for example `archive/notes/foo.md` becomes `notes/foo.md`).
The search index is updated immediately so the restored file is searchable again.
Refuses to overwrite if a file already exists at the original path, and rejects directories, missing files, or paths not inside `archive/`.
Paths are relative to the library root (MEMDRIVE_ROOT). Absolute paths are accepted only when they stay inside the root. The root is a hard sandbox: "../" escapes and symlinks pointing outside are rejected.

Input parameters:

- `path` (string, required): Archived file to restore, relative to the library root — must start with "archive/" (for example "archive/notes/2026-10-28.md").

### `plan_cleanup` (~152 tokens)

Generate a read-only cleanup plan for the library (two-step confirmation). Scans the library and proposes file actions — archiving empty files, duplicate files and orphan attachments — but changes nothing.
Returns a plan ID and a human-readable preview. The plan is also saved as JSON in the per-library cache dir (plan-<id>.json) for the `apply_plan` tool or the `memdrive-mcp apply <plan-id>` command.
Empty files, duplicates and orphan attachments are all proposed for archiving (recoverable); nothing is ever permanently deleted. Broken image references are listed as notes, not actions. Similar content is never proposed — only exact SHA-256 duplicates.
This tool never modifies any file.

### `apply_plan` (~288 tokens)

Execute a cleanup plan created by plan_cleanup, by its plan ID. Archives each action in the plan (empty files, duplicate files, orphan attachments) after re-verifying it; nothing is ever permanently deleted.
Before executing, each action is re-verified against the library: the path must still exist, the SHA-256 and modification time must match the plan, and the reason must still hold (empty file still empty, duplicate group still intact, orphan still unreferenced). Actions that no longer hold are skipped with the reason, and the rest still execute.
Plans older than 7 days are rejected. A plan ID that does not exist or is malformed is reported, and nothing is changed. The search index is refreshed incrementally after execution, so the next search reflects the change. A plan that has been executed is retired; running it again reports that it was already applied.
Use this when the user asks to execute a plan in chat (for example "执行这个清理计划"). The command-line `memdrive-mcp apply <plan-id>` does the same thing from a terminal.
Never call this tool automatically right after plan_cleanup. Always show the user the plan preview and plan ID first and wait for their explicit confirmation in the chat before calling.

Input parameters:

- `plan_id` (string, required): Plan ID returned by plan_cleanup (for example "20261114054412-c4ad79").

### `read_file` (~661 tokens)

Read a file from the MemDrive memory library.
TOTAL RESPONSE BUDGET — THREE LIMITS ON THREE DIFFERENT LAYERS. (1) HOST: the MCP host rejects a whole tool result over about 1 MB ("Tool result is too large. Maximum size is 1MB"), a host limit outside this server and separate from the model context window. (2) SERVER: text bytes plus base64 image bytes in one response must fit a single total budget — default 950,000 bytes, overridable with MEMDRIVE_RESPONSE_BUDGET_BYTES (clamped to 100,000–50,000,000). Text is guaranteed first; images share what remains and are downscaled in memory (trying higher JPEG quality before shrinking dimensions) or skipped with a note so a read never fails over an image and the file on disk is never modified. read_file returns a preview that fits the budget; to view the original image on disk, use open_image (only when the user asks). (3) MODEL: the model converts image pixels into tokens under its own limits, unrelated to the byte limits above.
TWO-STEP READING — USE THIS WHENEVER A DOCUMENT REFERENCES SEVERAL IMAGES, OR ONE LARGE IMAGE: Step 1, call read_file with include_images=false to read the text and get the list of referenced image paths. Step 2, call read_file once per referenced image, passing that image path itself as path, so the image alone gets the whole response budget and keeps the best quality the budget allows. Reading an image file directly is also limited by the same total budget: a single image is returned untouched when it fits, and downscaled in memory only when the image itself exceeds the budget.
Text files: returned as text (truncated at the smaller of the 1 MB text cap and the total budget, with a note when truncated).
Markdown files: with include_images=true (the default), the text is returned TOGETHER WITH every locally-referenced image that fits the remaining budget in the same response — each image becomes an image content block, so the model sees words and pictures together. Remote (http) and data: images are s…

Input parameters:

- `include_images` (boolean): For Markdown files: true (default) returns the text together with its referenced images, within the total response budget; false returns text only plus the list of referenced image paths (step 1 of t…
- `path` (string, required): File to read, relative to the library root (for example "notes/2026-10-28.md").

### `open_image` (~306 tokens)

Open an image file from the MemDrive memory library in the operating system's default image viewer (Photos on Windows/macOS, the default viewer on Linux), at full original quality.
USE ONLY WHEN THE USER EXPLICITLY ASKS TO VIEW THE ORIGINAL IMAGE — for example "open this image", "show me the original", or "在系统里打开这张图". NEVER call open_image automatically after a read_file, a save_file, or any other tool call, and never open an image the user has not asked to view. Opening a viewer window is a visible side effect on the user's screen; the model must not cause it unprompted.
The image itself is never returned to the model: the result is only a short text confirmation (or a note when no viewer is available, for example on a headless machine). To see the picture in the conversation, use read_file on the image path instead.
Only image files inside the library root can be opened. Non-image files, directories, missing files, and paths escaping the root are rejected with a note. The file on disk is never modified.
Paths are relative to the library root (MEMDRIVE_ROOT). Absolute paths are accepted only when they stay inside the root.

Input parameters:

- `path` (string, required): Image file to open, relative to the library root (for example "notes/assets/shot-20261028-120000-000.png"). Only call this when the user explicitly asked to view the original image.

### `open_containing_folder` (~357 tokens)

Open the folder containing a file or directory from the MemDrive memory library in the operating system's file manager (Explorer on Windows, Finder on macOS, the default file manager on Linux). A file opens its parent folder; a folder opens itself; an empty path opens the whole library root. The folder is opened without highlighting or selecting any file.
USE ONLY WHEN THE USER EXPLICITLY ASKS TO OPEN THE FOLDER — for example "open the folder", "show me this note in the file manager", or "在系统里打开这个笔记所在的文件夹". NEVER call open_containing_folder automatically after a read_file, a save_file, or any other tool call, and never open a folder the user has not asked to open. Opening a file manager window is a visible side effect on the user's screen; the model must not cause it unprompted.
The folder contents are never returned to the model: the result is only a short text confirmation naming the folder that was opened (or a note when no file manager is available, for example on a headless machine). To see what is in the folder, use list_directory instead.
Only paths inside the library root can be opened. Missing paths and paths escaping the root are rejected with a note. Nothing is modified.
Paths are relative to the library root (MEMDRIVE_ROOT). Absolute paths are accepted only when they stay inside the root.

Input parameters:

- `path` (string): File or folder whose containing folder to open, relative to the library root (for example "notes/2026-10-28.md" opens "notes"). Omit or pass "" to open the library root. Only call this when the user…

## Diagnostics

Captured diagnostic sections: Provenance, Vulnerabilities, Dependencies. The full working is on the page: https://verifymcp.io/servers/newmannn91-memdrive-mcp/memdrive-mcp#diagnostics

## Score history

- 2026-11-21: 70
- 2026-11-20: 69
- 2026-11-19: 68
- 2026-11-18: 68
- 2026-11-17: 68
- 2026-11-16: 59
- 2026-11-15: 67
- 2026-11-14: 66
- 2026-11-13: 67
- 2026-11-12: 67
- 2026-11-11: 51
- 2026-11-10: 66
- 2026-11-09: 65

## Common questions

### What is the io.github.Newmannn91/memdrive-mcp server?

io.github.Newmannn91/memdrive-mcp is listed in the public MCP registry as io.github.Newmannn91/memdrive-mcp. Local-first MCP memory: Markdown notes and images with full-text search for Claude and Cursor. This page covers its npm package (memdrive-mcp).

### Is the io.github.Newmannn91/memdrive-mcp server safe to use?

io.github.Newmannn91/memdrive-mcp scores 70 out of 100 on VerifyMCP. We recorded 1 known advisory against it as of 21 November 2026. It declares no install or post-install scripts. 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 io.github.Newmannn91/memdrive-mcp server expose?

io.github.Newmannn91/memdrive-mcp exposes 11 tools: list_directory, search_files, save_file, edit_tags, archive_file, and 6 more. Their descriptions and schemas cost roughly 4,370 tokens of context every time the server is loaded.

### Is the io.github.Newmannn91/memdrive-mcp server still maintained?

io.github.Newmannn91/memdrive-mcp is still listed as active in the MCP registry. We last reached this channel on 21 November 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 io.github.Newmannn91/memdrive-mcp server under?

io.github.Newmannn91/memdrive-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.

### What did VerifyMCP test on the io.github.Newmannn91/memdrive-mcp server?

VerifyMCP ran 16 checks on io.github.Newmannn91/memdrive-mcp, covering Malware scan, Known CVEs, Install scripts, Dependency health and 12 more. Endpoint Security and Transport & Reachability do not apply to an npm package, so they were not tested. The score does not measure everything, and the scoring documentation lists what it leaves out.

## Links

- npm package: https://www.npmjs.com/package/memdrive-mcp
- Socket report: https://socket.dev/npm/package/memdrive-mcp
- Repository: https://github.com/Newmannn91/MemDrive
- Changelog RSS feed: https://verifymcp.io/servers/newmannn91-memdrive-mcp/memdrive-mcp.xml
- Changelog JSON feed: https://verifymcp.io/servers/newmannn91-memdrive-mcp/memdrive-mcp/changelog.json
- Trust report (JSON): https://verifymcp.io/servers/newmannn91-memdrive-mcp/memdrive-mcp.json
- HTML version of this page: https://verifymcp.io/servers/newmannn91-memdrive-mcp/memdrive-mcp
