Use Taskhold with Claude Desktop
By Jake Marsh, who builds Taskhold. Updated 2026-10-26. Facts about other tools checked 2026-10-26. 3 min read.
Claude Desktop and claude.ai aren't coding agents, but they are a good place to plan: talk through a feature, then have Claude file it as Taskhold tasks your coding agents pick up. Anything that runs shows on the live board.
For the coding side, set up Claude Code; to plan from a spec, see break a spec into agent tasks.
What you get with Claude Desktop and Taskhold
- Claude can read your queue, file new tasks and split them into steps from a chat.
- Ask "what are my agents doing?" and Claude reads the tasks in progress with their steps and last messages.
- Reply to a blocked agent from the chat with a note on its task.
Connect Claude Desktop to Taskhold
- Start the 7-day trial, open Settings > API keys and create a key for this machine. It starts with atd_sk_. One key per tool and machine lets you revoke each one on its own.
- Open Customize > Connectors, click + Add, then Add custom connector.
- Enter the URL below. Under Authentication choose No sign in, and under Request headers add Authorization with the value Bearer and your key.
URL: https://taskhold.com/api/mcp
Header: Authorization: Bearer <your atd_sk_ key>To run it locally instead, add the hosted build to claude_desktop_config.json (Settings > Developer > Edit Config). Needs Node.js 18 or later:
{
"mcpServers": {
"taskhold": {
"command": "npx",
"args": ["-y", "https://taskhold.com/dl/taskhold-mcp-latest.tgz"],
"env": { "TASKHOLD_API_KEY": "<your atd_sk_ key>" }
}
}
}Tell Claude Desktop how to use the list
Claude Desktop has no rules file. Put the workflow in a Project's instructions so every chat in that project follows it:
## Tasks (Taskhold)
The person and I share a task list through the taskhold MCP tools. They watch a live board while I work.
- Session start: taskhold_get_current, then taskhold_updates for replies. Nothing in progress? taskhold_list status:todo and take the highest priority.
- taskhold_start before working on a task (agent: "Claude Desktop in <repo>").
- More than a couple of steps: taskhold_plan, then taskhold_start and taskhold_complete each step.
- taskhold_progress at milestones (N of M, or a percent) with a short message. Every few minutes at most.
- Need a decision: taskhold_block with one specific question, then check taskhold_updates.
- Done: taskhold_complete with what changed and how I checked it.
- A TODO or bug I'm not fixing now: taskhold_create (or taskhold_create_from_todo with file and line).One limit to know
On a Team or Enterprise plan, a connector with a fixed header gives everyone who uses it that one key, so every member would write into one Taskhold account. Use the connector on an individual plan, or the local config on your own machine.
What leaves your machine, and how to disconnect
Taskhold receives task titles, steps, progress messages, questions and summaries, whatever the agent or script sends. Taskhold never sends code or commands to your machine, and it doesn't read your repository.
To cut a tool off, open Settings > API keys and revoke its key. The next call fails with 401. Then remove the Taskhold entry from the tool's config so it stops trying.
See how far along your coding agents are.
$12/month or $120/year after the trial. Card up front. Cancel before day 7 and you pay nothing. Every agent you run is included.
Sources
- Claude Help Center: custom connectors using remote MCP, checked 2026-10-26
- Claude docs: authenticate a custom connector with request headers, checked 2026-10-26
- MCP docs: connect to local servers, checked 2026-10-26
Related
Updated 2026-10-26. Something wrong or out of date? Email hello@greatwork.company.