queued 2026-09-24 18:57:59 · started 18:58:00 · finished 18:58:02 · duration 2s · target v2.1.0 · chromium · trigger: cli-adopted · batch: 8 tests (terminal)
Activity replay 1 frames, captured only while the page was actually changing — idle waits are skipped automatically
+0.00s
Step timings follows the replay — click a step to see it there
| at | step | took |
|---|---|---|
| +0.75s | goto http://127.0.0.1:8000/demo/freight/ | 251 ms |
| +1.00s | wait_for_load_state networkidle | 406 ms |
| +1.41s | screenshot | 168 ms |
Times are when each step began, on the replay's clock. Add your own phases with
with ctx.timed("label"):.
Screenshots
Artifacts
| Run log | run.log |
| Playwright trace | trace.zip
open with: playwright show-trace trace.zip |
| Result JSON | result.json |
| Folder | results/DEMO-017/20260924-185800-r306/ |
Full log
Read from the S3 bucket (the last 20 KB; the whole file is the Run log link above).
[harness] test=DEMO-017 browser=chromium headed=False timeout=120s [harness] target: http://127.0.0.1:8000/demo/ [test] [ 0.75s] starting DEMO-017 [timing] goto http://127.0.0.1:8000/demo/freight/: 251ms [timing] wait_for_load_state networkidle: 406ms [timing] screenshot: 168ms [test] [ 2.21s] similarity 0.9999; 0.02% of pixels differ [test] [ 2.24s] PASSED [harness] security: 1 high, 3 medium, 1 low [harness] saved trace.zip [harness] result: passed
Security observations what the browser noticed while this ran — not a penetration test
| Severity | Finding | What to do |
|---|---|---|
| high | The page requested resources from other hosts Requests went to: analytics.acme-metrics.invalid (1). On an air-gapped network these cannot succeed -- so the page is either broken or waiting on a timeout -- and each one is a path by which data could leave the enclave. |
Vendor the resource locally, or remove the reference. This is worth chasing even when the page still appears to work. |
| medium | The application is served over plain HTTP http://127.0.0.1:8000/demo/ is not HTTPS, so anything typed into it travels the network in the clear and can be modified in transit. |
On an isolated lab network this may be a deliberate choice -- if so, note it and move on. Anywhere else, terminate TLS in front of the application. |
| medium | Missing Content-Security-Policy Without it, an injected script can run with the page's full privileges. This is the single most effective header against XSS. |
Set the Content-Security-Policy response header on the application or its reverse proxy. |
| medium | Missing X-Frame-Options (or CSP frame-ancestors) Without it, another site can embed this page invisibly and trick a user into clicking things (clickjacking). |
Set the X-Frame-Options (or CSP frame-ancestors) response header on the application or its reverse proxy. |
| low | Response advertises software version (server) server: pw-testhub -- this tells an attacker exactly what to look up known vulnerabilities for. |
Suppress or blank the server header at the proxy. |
| info | Cookie 'csrftoken' is readable by JavaScript (expected) This looks like a CSRF token, which the page's own JavaScript has to read in order to send it back. HttpOnly would break that. Nothing to do -- listed so you know it was checked and not missed. |
No action. Do check that SESSION cookies are HttpOnly. |
All of these appear together on the Security page.