- Home
- Skills
- APIs & Backend
- ClickUp Scribe
Works with the AI Agents you already use
ClickUp Scribe
Bulk read and write ClickUp boards over the REST API without burning the ClickUp MCP's daily quota - idempotent, dry-run first.
$5
ClickUp Scribe
Example session with this skill installed
I need to sync my local roadmap.json to the 'Product Development' list. Run a dry run first to see what tasks would be created.
- Read your context and instructions
- Compiled the clickup scribe
Dry run complete: 4 tasks would be created, 2 already exist, nothing was written.
I copied the bundled ClickUp Scribe scripts into a temporary folder and compared roadmap.json with the Product Development list over the REST API, so none of your ClickUp MCP quota was used.
$ scribe.py apply roadmap.json
list 901600000002: 2 tasks present
spec: 6 entries, 4 to create, 0 to update
= API Authentication Flow (completed)
+ OAuth2 Integration (to do)
+ Webhook Listener (to do)
= Database Schema Design (completed)
+ Rate Limit Middleware (to do)
+ Documentation Site (to do)
Dry run. Re-run with --apply to write.
| Action | Task | Type | Status from your spec |
|---|---|---|---|
| Exists, left as is | API Authentication Flow | Task | completed |
| Create | OAuth2 Integration | Subtask of API Authentication Flow | to do |
| Create | Webhook Listener | Subtask of API Authentication Flow | to do |
| Exists, left as is | Database Schema Design | Task | completed |
| Create | Rate Limit Middleware | Task | to do |
| Create | Documentation Site | Task | to do |
Existing tasks are matched by name across the whole list, so a re-run never creates duplicates. Nothing is ever deleted.
One mismatch to know about: your spec marks API Authentication Flow as completed, but on the board it is still
in progress. A plain apply leaves existing tasks untouched. Adding --reconcile would update it as well:
$ scribe.py apply roadmap.json --reconcile
list 901600000002: 2 tasks present
spec: 6 entries, 4 to create, 1 to update
= API Authentication Flow (completed)
+ OAuth2 Integration (to do)
+ Webhook Listener (to do)
= Database Schema Design (completed)
+ Rate Limit Middleware (to do)
+ Documentation Site (to do)
~ API Authentication Flow -> status=completed
Dry run. Re-run with --apply to write.
Next steps
- Check that OAuth2 Integration and Webhook Listener belong under API Authentication Flow.
- Say
"apply" to create the 4 tasks (--apply), or
"apply with reconcile" to also mark API Authentication Flow completed.
- After applying, running the same spec again plans
0 to create, so it is safe to repeat.
Connects securely to your tools. The creator never sees your data.
About this skill
The ClickUp MCP server has a small daily quota. One workspace walk or a 40-task import can exhaust it, and then reads and writes are blocked for about 23 hours. ClickUp Scribe gives your agent the REST API instead (100+ requests a minute) with the safety rails a bulk job needs.
Write
- Create, update, or delete one task with every common field (status, priority, due/start dates, estimate, tags, assignees by id, email, or username)
- Build a whole parent/subtask tree from one declarative JSON spec
--reconcileconverges an existing board onto the spec: creates what is missing, updates only the fields the spec declares- Create folders and lists, skipping any that already exist
Read
- Walk the full hierarchy, including folderless lists that naive traversals silently drop
- Every task and subtask on a list, re-nested, following pagination to the end
- Name search, single-task detail, and
export --as-specto clone a board's shape onto another list
Link git commits to tasks
- ClickUp's API has no endpoint for the GitHub/Bitbucket panel, so
commitsposts one comment per task listing short SHA, subject, and a web link, deduped by full SHA --rollupadds a per-subtask summary on the parent task
Safe by design
- Idempotent: tasks are matched by name across the whole list before anything is created, so a re-run creates only what is missing
- Dry-run first: nothing is written without
--apply - Statuses are validated against the list before the first write
- Serialized, paced, and every command reports how many API calls it spent
- Your token is never typed into chat: you store it once in your own terminal with a hidden prompt, and it is saved owner-only
Ships as stdlib-only Python 3 plus a Windows PowerShell 5.1 twin of every command, so it runs on Linux, macOS, or Windows with nothing to install.
Find more skills like this on the Claude Skills Marketplace, the Codex Skills Marketplace and the Gemini Skills Marketplace, or browse Claude Skills for APIs & Backend.
Deliberately left to the ClickUp MCP or UI, which do these in one call each: comments (other than commit links), docs, chat, time tracking, attachments, creating tags, dependencies, moving or merging tasks, reminders, custom field values, and full-text search. Descriptions and assignees are not reconciled. Tasks are matched by name, so renaming in a spec creates a new task (use set to rename).
How to use
- 1
Connect your agent to the Agensi MCP
Copy the prompt and paste it into Claude, ChatGPT, Cursor or any agent that supports MCP. It sets up the connection for you. Manual setup
- 2
Sign in to Agensi
Your agent shows a sign-in link. Sign in and click Allow.
- 3
Ask for what you need
Your agent loads this skill and uses it on your task. It always asks before buying a skill.
Reviews
No reviews yet
Be one of the first to try it. Every listed skill goes through the trust checks below.
Security scanned
Checked by our 8-point scan before listing
Fresh listing
Recently published to Agensi
Money-back guarantee
Doesn't do what its page promises? You get your money back.
Trust & safety
Security scan
Security scan passed 4 days ago
- Accounts
Clickup
- Software
Python 3 (Linux, macOS, or Windows)
Git
Permissions
Allowed Hosts