Skip to content
Back to skills

Clickup Debug Bundle

ASecurity

Collect a minimal redacted ClickUp connectivity and contract bundle without task bodies, tokens, or webhook secrets. Use when escalating a ClickUp API incident. Trigger with "ClickUp debug bundle", "ClickUp diagnostics", or "collect ClickUp evidence".

  • 2,719 stars
  • 0 votes
  • 0 copies
  • 3 views
  • Added September 21, 2026
ai-agentsgotestingapidocumentation

Works with

  • claude code
  • cli
  • api

Security analysis

A100/100

Pro scans all 2 files and shows the line behind each finding

Scanned September 21, 2026

npx -y skills add jeremylongshore/claude-code-plugins-plus --skill clickup-debug-bundle --agent claude-code

Installs into .claude/skills of the current project.

Are you the author of Clickup Debug Bundle?

Add the live security badge to your README. It updates with every re-scan.

Security grade badge for Clickup Debug Bundle
[![Security: A — Skills Directory](https://www.skillsdirectory.com/api/skills/jeremylongshore-clickup-debug-bundle-e018f57d/badge)](https://www.skillsdirectory.com/skills/jeremylongshore-clickup-debug-bundle-e018f57d)

More formats (shields.io, HTML) on the badges page. Keep it an A: scan every change in CI with Pro.

Download with Pro
SKILL.md
---
name: clickup-debug-bundle
description: >-
  Collect a minimal redacted ClickUp connectivity and contract bundle without task bodies, tokens, or webhook secrets. Use when escalating a ClickUp API incident. Trigger with "ClickUp debug bundle", "ClickUp diagnostics", or "collect ClickUp evidence".
argument-hint: "[incident-id] [output-directory]"
allowed-tools: Read, Glob, Grep, WebFetch, Write, Edit
version: 1.8.0
author: Jeremy Longshore <jeremy@intentsolutions.io>
license: MIT
tags:
- saas
- clickup
- diagnostics
model: inherit
effort: high
compatibility: Designed for Claude Code; live collection requires authorized ClickUp access and local retention controls
---
# ClickUp Redacted Diagnostic Bundle

## Overview

Create supportable evidence while preventing a troubleshooting artifact from becoming a second sensitive-data store. Make every included field explainable to the named reviewer.

## Prerequisites

- An incident identifier, approved output directory, retention deadline, and named reviewer
- A scoped credential stored outside command history and artifact content
- An allow-list of read-only probes and fields

## Tool Discipline

Use `Read`, `Glob`, and `Grep` to inspect the repository, adapters, configuration names, tests, and evidence. Use `WebFetch` only for current official ClickUp documentation. Use `Write` or `Edit` after confirming the target file, Workspace boundary, and requested mode.

## Current Contract

- Useful evidence includes status, endpoint version, latency, response type, sanitized error code, and rate headers.
- `GET /api/v2/user` and `GET /api/v2/team` can prove identity and authorized Workspaces without reading task content.
- Provider status is separate from tenant authorization and application behavior.
- Authorization headers, response bodies, user emails, task names, comments, attachments, and webhook secrets are excluded.

## Authentication

Use a personal token only for accountable individual/testing work or OAuth Authorization Code for a user-facing integration. Inject the token server-side through a governed secret reference, send it in `Authorization`, verify authorized Workspace IDs, and never print the token, OAuth client secret, or webhook secret.

## Instructions

1. Confirm incident scope, collection owner, field allow-list, output path, and deletion deadline.
2. Inspect local configuration names and versions without printing values.
3. Run bounded read-only identity/workspace and status probes; capture only approved metadata.
4. Record sanitized rate headers, status codes, timings, schema digests, and local adapter versions.
5. Scan the bundle for credential and content patterns; require human review before sharing.
6. Hash, transfer through the approved channel, and delete or retain on schedule.

## Approval Boundaries

Do not add raw headers, full environment dumps, task payloads, member records, or attachments to make the bundle more convenient.

## Output

Return bundle path, SHA-256, included field classes, excluded classes, scan result, reviewer, sharing decision, and deletion deadline.

## Error Handling

| Condition | Response |
|---|---|
| Redaction scan finds a secret | Do not share; quarantine, rotate, and regenerate. |
| Probe would read work content | Skip it and record the evidence gap. |
| Output directory is uncontrolled | Refuse collection until a governed location exists. |
| Incident scope expands | Obtain a new collection approval. |

## Examples

The example below is a redacted operator receipt; it contains no task text, member data, credential, or webhook secret.

```text
incident=CU-142; probes=3; task-bodies=0; secrets=0; sha256=recorded; review=approved; delete-by=2026-09-17
```

## Resources

- [Skill-specific official documentation](references/official-docs.md)
- [ClickUp status](https://status.clickup.com/)
- [Authentication](https://developer.clickup.com/docs/authentication)
- [Rate limits](https://developer.clickup.com/docs/rate-limits)

Files in this skill

  • SKILL.md3.9 KB
  • references/official-docs.md2.3 KB

Attribution

Is this your skill, or is something wrong with this listing? Request removal or report an issue. Author removals are honored within 72 hours.

Comments

Loading comments…

Newsletter

Get the weekly list of safe new skills, and the ones we pulled.