Skip to main content

How we check tools

Every check, when it runs, and what happens when one fails

Transparency first

Checked before it's listed, and every day it's live

Every tool on FolioNames goes through the same checks, whoever made it. Here is each one: what it looks at, when it runs, and what happens if a tool fails it.

  • Once a day

    Uptime check on every live tool

  • Every 6 hours

    Watch for new scripts, frames and forms

  • Every night

    Security scan, safety checks and speed test

  • Before listing

    Review, risk level and Google Safe Browsing

The checks, in order

From the moment a tool is submitted to every day it stays on the site.

Stage 1

Before it's listed

  1. Automated reviewMakers' tools

    A Claude-powered review checks the tool is legitimate and that what it asks for matches what it claims, and looks for phishing or misleading claims. A person looks at anything it flags.

  2. Editor checkTools we add

    An editor checks the tools our team adds from their public websites, and confirms the key facts against the tool's own pages.

  3. Risk levelEvery tool

    Each tool is graded Minimal, Limited, Elevated, High or Unacceptable. Sensitive permissions such as camera, microphone, location and payment raise the level.

  4. Google Safe Browsing and HTTPSEvery tool

    The site must use HTTPS and must not be on Google's list of unsafe sites.

Stage 2

Every day it's live

  1. UptimeOnce a day

    We check the site answers. Each night the last 30 days of checks become its Reliability score.

  2. Page changesEvery 6 hours

    We watch for scripts, frames or form destinations from websites no review has seen.

  3. Security scanNightly

    HTTPS, key security headers, and any exposed .git or .env file.

  4. Safety checksNightly

    Privacy policy, terms, contact details, trackers, forms, security headers and domain age (see below).

  5. Speed and accessibilityNightly

    Google PageSpeed measures how fast the site loads and how usable it is for everyone.

  6. Google Safe BrowsingNightly

    Every live site is checked against Google's list of unsafe sites again.

Stage 3

When something fails

  1. Down for too long6 checks in a row

    A tool that fails 6 uptime checks in a row is suspended (new listings get a 24-hour grace period). It comes back on its own after 3 healthy checks.

  2. Serious security finding2 scans in a row

    A high-severity finding on two nightly scans in a row suspends the tool until it's fixed. Missing headers alone never do.

  3. Code from a new websiteSeen twice

    A new script, frame or form destination seen on two checks in a row takes the tool off the site until it's reviewed again.

  4. Flagged as unsafeNext check

    A Google Safe Browsing flag fails the safety checks and sets the tool's Safety score to 0.

  5. A fault on our sideOver a fifth down

    If more than a fifth of tools look down at the same time, we assume the problem is ours and suspend nothing.

How we score apps: the FolioScore

Every listing gets one score out of 100, shown in a circle on its card and page. It's the average of five parts:

Speed

Google PageSpeed

How fast the site loads on a phone.

Accessibility

Google PageSpeed

How usable it is for everyone, including people using screen readers or keyboards.

Reliability

Our uptime checks

The share of our 15-minute checks it answered over the last 30 days, shown once it has a day of checks.

Ratings

Reviews

What signed-in reviewers say, weighted so one review can't swing it.

Safety

Our safety checks

The safety checks below, averaged with the AI risk review where there is one. A Google Safe Browsing flag makes it 0.

  • How many people use a tool isn't part of the score, so new tools aren't marked down.
  • A score needs at least three parts measured. Until then the circle shows a dash, and hovering over it says why.
  • For a tool's first week the circle has a dashed ring: the score is provisional.
  • Hover over any circle to see each part.

Safety checks: the same for every app

Every live listing is checked each night, and as soon as it's published, using only evidence anyone could repeat. No AI and no paid service is involved, so every tool is treated the same, whoever made it.

Points for each check

  • Security headers18 points

    A Content Security Policy (8), HSTS (6) and nosniff (4).

  • Privacy policy16 points

    Found from the page's links or its usual address.

  • Domain age12 points

    From the public registration record: full marks at two years or more.

  • No advertising trackers10 points

    2.5 points off for each tracker the page loads.

  • Forms10 points

    Sending data anywhere other than the site or a known payment, sign-in or form service costs points; a password sent elsewhere costs them all.

  • Terms of use8 points

    Found from the page's links or their usual address.

  • Contact details8 points

    An email address, phone number, or contact or about page.

  • No insecure content6 points

    Nothing loaded over plain http on a secure page.

  • Domain ownership6 points

    The maker has proved they own the domain. It doesn't apply to a listing no maker has claimed yet, so it neither helps nor hurts that listing's score.

The score is the share of the points that could be measured, so a check that couldn't run neither helps nor hurts. At least 40 points' worth must be measurable for a score.

Fails outright

  • A Google Safe Browsing flag
  • No HTTPS
  • An exposed .env or .git file

A higher bar for sensitive tools

For health, legal, finance, education, government and politics tools, the privacy policy, terms and contact details count half again, because mistakes there cost people most.

What our checks can't tell you

  • They don't test every feature inside a tool, or judge how good its answers are.
  • They can't see behind a sign-in, so paid areas aren't checked.
  • A good score isn't a guarantee: read a tool's own terms before sharing personal or sensitive information.
  • How popular a tool is never changes its score.

Seen a problem?

Use Report on any listing to tell us about a broken, misleading or unsafe tool, and we'll look at it. For a security issue in FolioNames itself, see Reporting a vulnerability below.

More detail

Automated review before publication

Every submission from a developer goes through an automated, Claude-powered review before it can go public. The reviewer checks that the app is legitimate, that the permissions it requests match what it claims to do, and flags red flags such as phishing or misleading claims. Listings our team adds from a tool’s public website are curated and safety-checked by an editor before they are published: the site is checked against Google Safe Browsing, must use HTTPS, and its key facts are confirmed against its own pages. Everything a developer submits is treated strictly as untrusted content — the reviewer never acts on instructions hidden inside a listing.

Five-level risk classification

In parallel, every app is graded on a five-level risk scale: Minimal, Limited, Elevated, High and Unacceptable. Sensitive permissions — camera, microphone, location and payment — automatically raise the level, and the classifier can only ever raise the code-derived risk, never lower it (so a cleverly worded description can't talk its way down). A developer's app is published automatically only when the review recommends it, the risk is Minimal, Google Safe Browsing finds nothing and the developer has verified the domain. Anything else, or anything flagged, is routed to a human moderator or rejected with feedback.

How tools run — sandboxed by default

A tool that can be tried on the page runs inside a sandboxed iframe. The sandbox lets it run scripts and forms but gives it a temporary, isolated origin rather than its own or FolioNames', so it cannot read the FolioNames page around it, your FolioNames session or any other tool. The same isolation means the tool cannot use its own cookies or saved sign-in inside the frame, so some tools only work fully when opened in a new tab. What you type or upload into the tool itself still goes to that tool's maker, under their own terms and privacy policy. Before a tool is listed we check whether it can be safely embedded (its X-Frame-Options and CSP frame-ancestors headers). Tools that can't be embedded open in a new tab instead, rather than being forced into the sandbox.

Safe handling of submitted URLs

Every URL a developer submits — and every URL we later health-check — is fetched through a guarded client that blocks requests to private networks, loopback and cloud-metadata addresses, rejects non-standard IP encodings, and limits redirects. This means submission and monitoring can never be turned into a tool for probing internal infrastructure (SSRF).

Input sanitisation & XSS protection

All text a developer enters — name, tagline, descriptions — is sanitised server-side with DOMPurify before it is stored. Rendered rich text is sanitised the same way, and structured data (JSON-LD) is escaped, so a listing can never inject scripts into the pages that display it.

Trust, verification & reviews

Where a trust score has been calculated, it is based on review outcomes, ratings, uptime and security history, and shown on the listing's health report. A listing without one simply shows none. Developers can optionally verify their identity to earn a verified badge — it builds trust, but isn't required to publish. Ratings and reviews come from signed-in users.

Your data

Imperium Business Intelligence Ltd is the data controller under UK GDPR and the Data Protection Act 2018. We do not sell your data; we share it only with the service providers needed to run the platform (such as hosting, email and error monitoring). Authentication is handled in-house with securely hashed passwords and optional two-factor authentication, and you can export or delete your account and data at any time. For full detail, see our Privacy Policy.

Reporting a vulnerability

Found a security issue? Please report it via our security.txt file or email andrew.hoyle@imperium-bi.co.uk. We aim to acknowledge within 24 hours and give an initial assessment within 72 hours, and we'll credit you (with your permission) once it is resolved. Please give us reasonable time to fix an issue before disclosing it publicly.