Atlas / MCP servers / kazuph / TaskManager

TaskManagerSAFE

mcp/kazuph/taskmanager
Verdict
SAFE
Grade
B
Trust score
89 /100
Exposed tools
10 7r · 2w · 1d
Transport
stdio
License
MIT
Stars
216
01

Overview

From the repository's own README, as read at the audited commit. Badges and raw HTML are left out.

Model Context Protocol server for Task Management. This allows Claude Desktop (or any MCP client) to manage and execute tasks in a queue-based system.

Quick Start (For Users)

Prerequisites

  • Node.js 18+ (install via brew install node)
  • Claude Desktop (install from https://claude.ai/desktop)

Configuration

  1. Open your Claude Desktop configuration file at:

~/Library/Application Support/Claude/claude_desktop_config.json

You can find this through the Claude Desktop menu:

  1. Open Claude Desktop
  2. Click Claude on the Mac menu bar
  3. Click "Settings"
  4. Click "Developer"
  1. Add the following to your configuration:
{
"tools": {
"taskmanager": {
"command": "npx",
"args": ["-y", "@kazuph/mcp-taskmanager"]
}
}
}

For Developers

Prerequisites

  • Node.js 18+ (install via brew install node)
  • Claude Desktop (install from https://claude.ai/desktop)
  • tsx (install via npm install -g tsx)

Installation

git clone https://github.com/kazuph/mcp-taskmanager.git
cd mcp-taskmanager
npm install
npm run build

Development Configuration

  1. Make sure Claude Desktop is installed and running.
  1. Install tsx globally if you haven't:
npm install -g tsx
# or
pnpm add -g tsx
  1. Modify your Claude Desktop config located at:

~/Library/Application Support/Claude/claude_desktop_config.json

Add the following to your MCP client's configuration:

{
"tools": {
"taskmanager": {
"args": ["tsx", "/path/to/mcp-taskmanager/index.ts"]
}
}
}

Available Operations

The TaskManager supports two main phases of operation:

Planning Phase

  • Accepts a task list (array of strings) from the user
  • Stores tasks internally as a queue
  • Returns an execution plan (task overview, task
Read from source at commit 30b6ee563032OBSERVED · 2026-10-06
02

Connect

Built from this server's own package name, version and transport as found in its source — not copied from anyone's documentation, so it cannot drift against a page we do not control.

claude-code
claude mcp add mcp-taskmanager -- npx -y @kazuph/[email protected]
claude-desktop
{
  "mcpServers": {
    "mcp-taskmanager": {
      "command": "npx",
      "args": [
        "-y",
        "@kazuph/[email protected]"
      ]
    }
  }
}
03

Exposed tools (10)

7 read · 2 write · 1 destructive. Blast radius: 1 tool can delete or overwrite — an agent that can be talked into calling a tool can be talked into calling this one.

ToolRiskDescription
add_tasks_to_requestwriteAdd new tasks to an existing request. This allows extending a request with additional tasks.\n\n
approve_request_completionreadAfter all tasks are done and approved, this tool finalizes the entire request. The user must call this to confirm that the request is fully completed.\n\n
approve_task_completionreadOnce the assistant has marked a task as done using
delete_taskdestructiveDelete a specific task from a request. Only uncompleted tasks can be deleted.\n\n
get_next_taskreadGiven a
list_requestsreadList all requests with their basic information and summary of tasks. This provides a quick overview of all requests in the system.
mark_task_donereadMark a given task as done after you
open_task_detailsreadGet details of a specific task by
request_planningreadRegister a new user request and plan its associated tasks. You must provide
update_taskwriteUpdate an existing task
04

Trust audit

SAFEgrade B · trust 89/100 Nothing in the source contradicts what it says it does. Grade A is reserved for packages that have also passed the behavioural sandbox.

LayerWhat it checksResult
L0Provenance & inventoryPASS
L1Static analysis of the codePASS
L2Instruction surface (what it tells the agent)PASS
L3Class-specific surfaceWARN
L4Behavioural (sandbox)SKIPPED

What the source does

Filesystem
declared (1 observation(s))
Network
none-observed
Shell
none-observed
Dependencies
not all pinned
Secrets in source
none-found

Findings (2)

MEDIUMFilesystem / path · mcp.destructive_tools · CWE-22, CWE-59
delete_task
Why it matters. 1 tool(s) can delete or overwrite
Fix. prefer a read-only mode or scoped tokens; the page states the blast radius
LOWSupply chain · supply.unpinned · CWE-829, CWE-1357
package.json
@modelcontextprotocol/sdk, zod, @types/node, shx, typescript, vitest
Why it matters. 6 dependency range(s) float
Fix. pin exact versions or ship a lockfile

Gates applied: no_behavioural_pass.

Audited 2026-10-06 · audit v0.4.1 · source sha 30b6ee563032full audit observations/trust-audit/mcp-server/kazuph__taskmanager.json · Report an issue / request a re-scan
05

Audit history

Every audit this server has had. A grade with a past is a grade somebody is still checking.

DateSourceVerdictGradeScoreChange
2026-10-0630b6ee563032SAFEB89first audit
06

Questions

What tools does TaskManager expose?

10 in total: 7 read-only, 2 that write, and 1 that can delete or overwrite (delete_task). Every one is listed on this page with its risk.

Is TaskManager safe to connect to an agent?

The audit found nothing in the source that contradicts what it says it does, and graded it B (89/100). Grade A is held back for packages that have also passed a sandboxed behavioural run, which is why a clean server reads B. Separately from the audit: 1 of its tools can destroy data, so scope the token you give it to what you actually need.

What credentials does TaskManager need?

No credential environment variables were found in its source, so it appears to need none.

How does TaskManager run?

It speaks stdio, so it runs as a local process your client starts. It is published on npm as @kazuph/mcp-taskmanager at 1.1.1.

How current is this page?

The grade is for one exact copy of the source (30b6ee563032), read on 2026-10-06. The repository is watched and re-audited when it changes.

Advertisement