One evidence trail. A clear handoff to your coding assistant.
demo.example/Frozen test report2 of 2 findings shown · Quick scan
Exported Quick scan and recheck of our own owner-authorized test site, shown under a fictional alias. Not customer data.
01Detect
See what needs attention.
Your pages do not set a script policy.
/Low severity
Your pages do not ask browsers to remember HTTPS.
/Low severity
0 of 2 header checks passed · 1 of 7 areas assessed
Coverage, not score. Unassessed areas remain unknown.
02Prepare a repair
Give your AI coding tool a scoped fix.
REPAIR BRIEF
Change
Review script policy and HTTPS-only visits on /.
Boundary
2 of 2 findings shown. Review this scope and preserve intended behavior.
Test
Test intended behavior. Review HTTPS support and rollback.
Verify
Deploy separately. Run a comparable recheck.
Review → apply → test
You stay in control. No tool or deployment is launched.
03Recheck
See what actually changed.
From a recorded recheck of our own test site
Quick scan → comparable recheck ·
Earlier evidence · /
Your pages do not set a script policy
Recorded recheck
No longer observed
Earlier evidence · /
Your pages do not ask browsers to remember HTTPS
Recorded recheck
No longer observed
Same checked settingsEarlier: 1 of 7 areas assessedRecheck: 1 of 7 areas assessed
0 of 2 → 2 of 2
Header checks passedConfiguration presence, not overall security.
Sample methodology and coverage
HTTP responsesAssessed
Public exposureNot in this sample
DNS and emailNot in this sample
TLSNot in this sample
Authenticated accessNot in this sample
RepositoryNot in this sample
Model analysisNot in this sample
3 of 3 scheduled configuration settings assessed. The sample has 2 completed weighted header checks. Earlier: 0 passed; recheck: 2 passed. A numeric score requires at least 5 completed weighted checks.
Technical header details
Request target: https://demo.example/
Content-Security-Policy: not returned
Finding ID: 4af59a3b7f93732e570d9456
Recheck: No longer observed
Strict-Transport-Security: not returned
Finding ID: 69345dca363b732777d97c71
Recheck: No longer observed
The owned fixture uses synthetic HTTP transport to represent HTTPS headers. The report pair was exported through the shared report contracts, then aliased and identity-redacted for this static presentation. The repair handoff illustrates the frozen guidance; it is not an exact exported brief. Missing headers alone do not establish exploitability. Intended access and application behavior were not tested. No overall safety claim is implied.
How it works
Request an invitation, then work from evidence to a repair and recheck.
01
Invitation beta
Request access
Ask for an invitation using the beta access link.
02
Invitation beta
Create an account
Create your account after your invitation is confirmed.
03
Invitation beta
Scan
Choose Quick or Deep and confirm your authority to assess the target.
04
Invitation beta
See the evidence
Read what was observed, why it matters and what was not assessed.
05
Invitation beta
Prepare a repair
Take a scoped brief to your coding tool. Review and test the changes.
06
Invitation beta
Recheck the deployment
After deployment, compare fresh observations on a compatible basis.
What you can check
Evidence, within an authorized scope.
Invitation beta
Your website
Understand the configuration gaps that need your attention.
Choose the depth of your website check
Profile
Quick
Deep adds
Needs
Your attestation
Everything in Quick, plus ownership proof
Covers
Security headers
HTTPS/TLS connection facts
Sampled same-origin scripts for secret patterns
Bounded same-origin crawl
Environment-file exposure check
Quick
Needs
Your attestation
Covers
Security headers
HTTPS/TLS connection facts
Sampled same-origin scripts for secret patterns
Deep adds
Needs
Everything in Quick, plus ownership proof
Covers
Bounded same-origin crawl
Environment-file exposure check
Not included: signed-in browser testing, form submissions or active exploitation.
Invitation beta · Limited access
Your repository
Bring concrete evidence from a pinned snapshot into your repair discussion.
Secret patterns
Dependency advisories
Focused CI configuration rules
Not included: code execution or complete source-code coverage.
Planned
Your team is building with AI. Who’s reviewing what gets shipped?
Application reviews connect the builder, business owner and reviewer around the intended use, recorded evidence and accountable human decisions.
Builder
Shares what was built and the evidence.
Business owner
Defines intended use and business context.
Reviewer
Records a scoped decision and rationale.
A recheck does not approve an application. The decision stays with a person.
A useful finding gives you evidence you can act on. See what was observed, what it may mean and where coverage stops.
Authorization comes first.
Scan what you’re authorized to assess. Quick scans need your attestation; Deep scans add ownership proof. Rechecks stay inside the same authorized scope.
Facts stay separate from inference.
Understand what was observed before deciding what to change. Facts, possible impact and unchecked areas stay separate. Unchecked is never shown as safe.
Coverage is not score.
A Configuration score appears only after at least 5 weighted checks complete, so a couple of headers can’t earn a grade. It summarizes checked settings, not overall security. A repair statement isn’t fresh evidence.
Coming next: agent & MCP configuration checks
Building with MCP servers and coding agents? We’re adding static checks for tool permissions, exposed secrets and risky configuration.
Not live yet
Turn security evidence into your next good change.