Scitor for agents
Your agents already know how to use Scitor
Tickets are GitHub issues. Replies are comments. Configuration is a file in your
repository. Any agent that can run gh can already run your support desk — the skill just teaches it the conventions.
$ npx skills add scitorapp/agent-skills Works with Claude Code, Codex, Cursor, Copilot, Devin, Gemini, and OpenCode.
Everything your team does, from the terminal
Because Scitor is built on GitHub's own primitives, there is no separate agent surface to maintain — and no feature gap between what a person can do and what an agent can do.
Read the queue
Tickets are GitHub Issues or Discussions, so listing, filtering, and reading them is plain gh. Scitor's AI labels — sentiment, priority, category, spam — become filters your agent can pivot on.
gh issue list --repo acme/support \
--label "sentiment:negative" \
--label "priority:urgent"Reply to customers
A reply is a comment that starts with /send. The Markdown below it becomes the HTML email that reaches your customer, threaded to the original conversation.
gh issue comment 482 --repo acme/support --body '/send
Hi Dana — the export now includes archived
records. Please try again.
Acme Support'Triage and route
Priority, assignment, SLA state, follow-ups, saved replies, and customer history are all slash commands. Anything an agent on your team can do from GitHub, a coding agent can do too.
/priority urgent
/assign @dana
/followup 2d Check the customer confirmedChange configuration
Routing rules, SLAs, auto-replies, and the knowledge base all live in one YAML file in your repository. Agents change behaviour by opening a pull request — reviewed like any other change.
.github/scitor.yamlAnd your product becomes agent-ready too
Every Scitor knowledge base ships an llms.txt
Your customers increasingly arrive with an agent already open. A Scitor-hosted knowledge base publishes machine-readable copies of itself automatically, so those agents answer from your documentation instead of guessing at it — no extra work from you.
- /llms.txt
- An index of every page in your knowledge base, following the llmstxt.org convention.
- /llms-full.txt
- Your entire knowledge base in a single file, for agents that would rather read everything once than follow links.
- /any-page.md
-
The raw Markdown behind any page — add
.mdto its URL. No HTML to parse, no navigation to strip out.
See it on our own docs: support.scitor.io/llms.txt
Nothing new to trust, install, or maintain
No API to learn
There is no Scitor SDK, REST client, or authentication flow to work through. If your agent can run gh, it is already integrated — nothing to install on our side.
Every action is auditable
An agent's reply is a GitHub comment with an author and a timestamp. Your team reviews agent work in the same place it reviews everything else, and revokes access by removing repository access.
Works with every agent
Claude Code, Codex, Cursor, Copilot, Devin, Gemini, OpenCode. Nothing here is agent-specific — it is the GitHub CLI and Markdown, so it keeps working as the tools change.
Want agents working the queue without you?
The skill is for agents you drive. Scitor also supports agents that run on their own inside GitHub Actions — triaging new tickets, drafting replies from your docs, and writing weekly reports, all with a full audit trail.
Explore agentic workflowsPoint your agent at your support queue
Install Scitor on a private repository, install the skill, and your existing agent setup can start working tickets today.