The Safeguard Firefox add-on is a sidebar for AI search against your account. It does not scan code, enforce policy or work without sign-in. Its Firefox listing declares data collection that the docs describe differently, and a known cookie quirk can ask you to sign in each session.
It is AI search only
The browser extension docs state that scope plainly. For findings inside your editor, they point to the IDE extensions. For gating a pull request, use the CI and repository integrations. For long reports, use the web app. The sidebar is for quick questions.
It needs an account
The listing says a Safeguard account is required. Without one there is nothing for the sidebar to answer from.
What does the listing say about data?
Read this section before you roll the add-on out.
On the Firefox listing, the permissions area shows optional access to data for all websites. The data collection area shows required categories declared by the developer: authentication information, website activity and website content.
The docs describe a narrower picture. They say the extension collects no data, makes no Safeguard API calls of its own and stores no credentials. For Chrome and Edge they say only the sidePanel permission is requested, with no host permissions. For Firefox they say the add-on declares that no data is collected.
We cannot resolve that gap from public pages. A sensible reading is that the sidebar loads Safeguard's own web page, which handles your session, and that the declarations on the listing reflect what that page processes. That is an interpretation, not a statement from Safeguard. If this matters for your review, ask Safeguard in writing.
What about page content?
The listing says the sidebar reads a page only when you attach the current tab, and clicks nothing. So the practical control is yours: do not attach pages that hold personal or customer data.
The session quirk
The docs say the web app's session cookie uses SameSite=Lax, which can stop login persisting across a browser restart inside the panel. The fix is server side. Until then, expect to approve the browser again in a new session. It is an inconvenience, not a security gap.
Browser support
The docs list Firefox 113 or later. Temporary add-ons loaded for development are removed when Firefox restarts, which is a note for developers, not for people who install from the store.
Reviews and track record
On our check the listing showed no ratings yet and version 1.0.4, updated 13 October 2026. A new listing has little public history, so pilot it with a few people first.
Is any of this a reason not to use it?
Not by itself. It is a reason to read before you install, which is true of any extension that can see a page. The docs describe a deliberately thin design, and the listing is specific about what the sidebar reads and when. Make the decision with your own risk rules, not ours.
What to do
- Read the data and permissions section yourself.
- Ask Safeguard to explain any difference from the docs in writing.
- Pilot with two or three people before a team rollout.
- Use a work profile so the add-on never sits in personal browsing.
- Keep the IDE extension and CI gate for the jobs a sidebar cannot do.
Quick check
Can the Firefox add-on enforce a security policy?
No. It is AI search in a sidebar. Enforcement belongs in your repository or pipeline.
Do the docs and the listing agree on data collection?
Not fully. The docs say no data is collected, while the listing declares authentication information, website activity and website content. Ask Safeguard if it matters for your review.
Sources
Last reviewed: 18 October 2026.
