Support admins built CLIs and skills for Confluence, Jira and Zendesk with Claude Code, and used a custom GPT to draft ticket answers that had to be pasted back manually.
Support admins built CLIs and skills for Confluence, Jira and Zendesk with Claude Code, and used a custom GPT to draft ticket answers that had to be pasted back manually.
One admin, quoted from Hacker News, connected Zendesk, Jira, Confluence and GitHub to Claude Code. They built a CLI for each of Confluence, Jira and Zendesk, with matching skills and an OAuth login for each, in a couple of days. A separate admin built a custom GPT that answers tickets. Its replies had to be copied back into Zendesk by hand.
The commenter says cross-tool search dropped from about 15 minutes of manual hunting to about one minute. The custom GPT drafts worked well, but the paste-back step stayed manual.
A Zendesk draft is stored locally and can't be written into the ticket from outside, so an external drafter can only post an internal note or feed a sidebar app.
I work on support tickets in [Zendesk/Jira/other helpdesk] and keep answers in [Confluence/knowledge base]. Build me a small command-line tool and reusable skill that pulls a ticket by ID, searches our docs for relevant articles, and drafts a reply in our tone. Do not post or send anything; show me the draft and ask for approval first. You're done when I can run it on [ticket ID] and get a draft with links to the sources.
Plan, execute and review a cross-cutting refactor in small batches, with Claude Code making the edits and GitHub Copilot reviewing each PR.
Export the Descript transcript, let ChatGPT pick the five strongest moments, and get 10 platform-ready posts in a review doc within minutes.
Give it a keyword and get a brief built from the pages that rank today: headings, gaps, questions to answer and a target length.