GitHub's Copilot app can now explain and fix your code scanning alerts
SummaryA new GitHub security extension, in public preview, puts a Security canvas inside the GitHub Copilot app: browse code scanning alerts, ask Copilot to explain or fix them, dismiss false positives, or review local changes before a pull request. Not tested by MODEL.
· MODEL Editors · 2 min read
Some links may be affiliate links. Disclosure

What it is. GitHub announced (opens in a new tab) on Nov. 15 a GitHub security extension for the GitHub Copilot app, now in public preview. It adds a canvas labeled Security next to your session, where you can work through a repository's code scanning alerts and ask the Copilot agent to find, explain and fix them, or dismiss them (GitHub Docs (opens in a new tab)).
How it works. You install the github-security plugin, either from the app (Customize, search "github-security", Install) or with copilot plugin install github-security@awesome-copilot in Copilot CLI. The canvas lists open code scanning alerts by severity. Explain and Fix buttons send the selected alerts to the agent in your session. A separate local review, started from the canvas or with /security-review, checks your in-progress changes for new vulnerabilities before you open a pull request.
What it needs. A Copilot plan and the GitHub Copilot app. On Copilot Business or Enterprise, the Copilot app policy must be enabled by an admin. To see existing alerts, code scanning has to be on for the repo, and private repos require GitHub Code Security. The local review works without code scanning or a Code Security license. GitHub.com and GitHub Enterprise Cloud with data residency are supported; GitHub Enterprise Server is not.
Who it's for. Developers who already use code scanning and keep a backlog of alerts, and teams that want a quick security pass on agent-written changes before review.
The catch. GitHub's docs say Fix only edits your local working copy: no commit, push or pull request, and the alert stays open until the fix lands on the branch and code scanning runs again. It covers code scanning alerts only, not Dependabot or secret scanning. Explain, Fix and local reviews use AI credits like other prompts. Dismissing an alert updates it on GitHub, and GitHub calls the local review a focused check of your changes, "not a full repository security audit or a replacement for code scanning."
The MODEL take. Moving the alert queue next to the agent that can fix it removes a lot of tab-switching. Treat each Fix as a draft: read the diff and run your tests, as GitHub itself advises.
Not tested by MODEL. Not sponsored. No affiliate links.
Get three tested tools every Friday
Free, weekly. Unsubscribe anytime.
Sources
Keep reading

Hot take
Google caught a fully automated AI attack. It didn't need a zero-day. It read the patch.
Google's latest AI Threat Tracker describes exploit_agent, a multi-agent system that diffs web software before and after each patch, drafts an exploit and runs it with no human in the loop. The same report counts hundreds of AI-written exploits that don't work. Our take: sloppy AI code now hits defenders from both sides.
· MODEL Editors

Hugging Face x OpenRouter
Hugging Face now routes its LLM calls through OpenRouter
Hugging Face switched Inference Providers from its own router to OpenRouter's. Same token, same API, a provider picked per request, and retries when one fails. One breaking change: your own provider keys no longer work for chat models. Not tested by MODEL.
· MODEL Editors

Builder tip
Amazon caught AI agents hunting exposed LLM gateways for API keys. Lock yours down.
Amazon's decoy sensors captured attacker AI agents' own campaign records: they scanned for exposed LiteLLM, One API, Dify and similar apps, pulled the provider keys inside, and used stolen keys to pay for more hunting. Every technique was old. Here's the checklist.
· MODEL Editors

xeet
xeet renders X posts to PNG without a browser
Guillermo Rauch's new open-source tool rebuilds X's post layout in Rust and WebAssembly, so apps, agents and CI jobs can preview a post as an image in milliseconds instead of screenshotting headless Chrome. Not tested by MODEL.
· MODEL Editors
FREE · EVERY WEEK
The top AI tools this week, tested.
Every week MODEL's editors pick the AI tools worth your time, try them hands-on, and tell you what's good, what's not, and who it's for. Free, in your inbox.
Free. Unsubscribe anytime. Our picks are editorial and can't be bought.