Tenable SC Agent — Security Center Vulnerability Analyst

SC Agent turns any Tenable Security Center vulnerability page into a conversation. It’s a Manifest V3 Chrome extension that injects a floating AI vulnerability analyst into Security Center: open a finding, click the shield in the corner, and ask anything. The agent answers using the vulnerability data scraped directly from the page — powered by the Anthropic API.

What it does

• Analyzes vulnerabilities in any language. Ask in English, Turkish, Spanish, Japanese — the agent reads the on-page finding (plugin, CVEs, CVSS, affected hosts, solution) and replies in the language you wrote in.
• Triggers your own automation. Define commands in an editable tools.md (triage, remediation, ticket, or anything you invent). Ask “triage this” or “open a ticket” and the agent runs that playbook against the live vulnerability data and returns structured, ready-to-use output.
• Grounds every answer in page context. The vulnerability title (from the h2.vuln-title element) and details (from every div.detail-section) are handed to the model automatically — no copy-pasting or prompting from scratch.
• Stays out of the way. A keyword-gated launcher (e.g. shows only on vulndetails pages) that expands from a compact modal to a full panel, running inside a Shadow DOM so it never clashes with Security Center’s own UI.

How it works

A content script detects matching Security Center pages, scrapes the visible vulnerability data, and streams the conversation to the Anthropic Messages API through the extension’s background service worker. The agent’s persona lives in an editable agent.md (default: Vulnerability Analyst) and its automation commands in tools.md, both configurable from the extension’s settings page. Bring your own Anthropic API key; everything runs locally in the browser.

See the repository README for install steps, configuration, and a demo video.