UserProof / Free tools / CSP report reader
CSP report reader: what your Content Security Policy blocked, and the fix. For founders who build with a coding agent
A CSP violation report tells you what a Content Security Policy blocked: the URL, the directive that stopped it and whether it was enforced or only reported. For founders who build with a coding agent: paste reports below, plus your policy if they don't carry it. Each blocked item comes back matched to its analytics or session replay tool, with the source to add, its directive, and your policy with the fixes in place.
Read a report
Example result for the reports in the box
4 reports read: 3 of the 11 tools named, 1 from browser extensions.
| Blocked | Directive | Tool | What to add |
|---|---|---|---|
https://region1.google-analytics.com | connect-src | Google Analytics 4 (Google tag) | https://*.google-analytics.com in connect-src (a new directive, seeded with the sources it fell back to) |
https://us-assets.i.posthog.com | script-src-elem | PostHog | https://*.posthog.com in script-src |
chrome-extension | script-src-elem | Not your site | Nothing. An extension in the visitor's browser injected it; ignore these. |
https://ingest.getuserproof.com/sdk.js | script-src-elem | UserProof | https://ingest.getuserproof.com in script-src |
Your policy with the fixes
default-src 'self'; script-src 'self' 'nonce-r4nd0m' https://www.googletagmanager.com https://*.posthog.com https://ingest.getuserproof.com; img-src 'self' data: https:; connect-src 'self' https://*.google-analytics.com https://www.googletagmanager.com https://*.google.com https://*.posthog.com https://ingest.getuserproof.com; worker-src 'self' 'nonce-r4nd0m' https://www.googletagmanager.com https://*.posthog.com https://ingest.getuserproof.com blob: data:; report-uri /csp-reportsAlso added, so the same tools don't fail on their next request: https://www.googletagmanager.com in connect-src for Google Analytics 4 (Google tag); https://*.google.com in connect-src for Google Analytics 4 (Google tag); https://*.posthog.com in connect-src for PostHog; blob: in worker-src for PostHog; data: in worker-src for PostHog; https://ingest.getuserproof.com in connect-src for UserProof.
To check the result before you ship it, paste it into the CSP checker.
How the reader decides
The blocked value. A URL on another host is matched against the hosts the 11 tools list on their own CSP pages: UserProof, Google Analytics 4, PostHog, Sentry, Hotjar, Microsoft Clarity, FullStory (US and EU), LogRocket, Mouseflow, Lucky Orange, Smartlook. A match gives the vendor's own source, such as https://*.posthog.com. A host that matches none of them gives its origin, and a URL on the page's own origin gives 'self'. inline is an inline script, event handler or style. eval is a string run as code. data and blob are those URL schemes. Anything from chrome-extension, moz-extension or safari-web-extension came from the visitor's browser, not your site.
The directive to edit. Browsers report the effective directive, such as script-src-elem, even when your policy only sets script-src. The reader walks the same fallback list the browser does (for scripts: script-src-elem, then script-src, then default-src) and edits the directive your policy actually has. When that is a fallback (default-src for a request, or script-src for a worker), it adds the directive itself, seeded with the fallback's sources, so nothing else in the page loses access and script-src never gets blob:.
The fixed policy. When it knows your policy, the reader adds each source and then runs the CSP checker's rules on the result. For each tool it named, any source that tool will need next (usually its data host in connect-src once its script loads) is added too and listed under the table.
The example, line by line
The box opens on four reports against one policy: default-src 'self'; script-src 'self' 'nonce-r4nd0m' https://www.googletagmanager.com; img-src 'self' data: https:; report-uri /csp-reports. Google Analytics' script loaded, but its hits to region1.google-analytics.com were refused: the policy has no connect-src, so default-src 'self' decides, and the fix is a new connect-src 'self' https://*.google-analytics.com. PostHog's script was blocked in script-src. The console line shows UserProof's script blocked the same way. The chrome-extension report is noise. The fixed policy adds both script hosts to script-src and the data hosts Google Analytics, PostHog and UserProof send to in connect-src. PostHog's recording worker gets its own worker-src with blob: and data:, seeded from script-src, so those don't fail next.
The three formats it reads
| Where it comes from | What it looks like | Fields the reader uses |
|---|---|---|
report-uri endpoint | POST with Content-Type application/csp-report, one object with a csp-report key | blocked-uri, effective-directive (or violated-directive), original-policy, disposition, document-uri, script-sample |
report-to endpoint (Reporting API) | POST with Content-Type application/reports+json, objects with "type": "csp-violation" and a body | body.blockedURL, body.effectiveDirective, body.originalPolicy, body.disposition, body.documentURL, body.sample |
| Browser console | Chrome: Refused to load the script '...' because it violates the following Content Security Policy directive: "..."; Firefox: The page's settings blocked ... | The verb (load the script, connect to, frame, create a worker from, execute inline script), the URL, and the directive Chrome quotes with its sources |
Field names from MDN's report-uri and report-to pages, read November 18, 2026. MDN notes that a blocked URL on another origin is cut to its scheme, host and port, and that script-sample carries the first 40 characters of inline code only when the directive has 'report-sample'.
How to start getting CSP reports
A policy reports only when it says where to send reports. Send both directives, because MDN says browsers that support report-to ignore report-uri, and report-to has only worked across current browsers since March 2026:
Reporting-Endpoints: csp="https://example.com/csp-reports"
Content-Security-Policy: default-src 'self'; script-src 'self' https://ingest.getuserproof.com; connect-src 'self' https://ingest.getuserproof.com; report-uri https://example.com/csp-reports; report-to csp
Neither directive works in a <meta> tag, so send the policy as a header. To test a change without breaking anything, send it as Content-Security-Policy-Report-Only first: browsers report what it would block and block nothing. Reports are written by visitors' browsers, so treat them as untrusted input at your endpoint, and expect extension noise.
How many startups hear what their CSP blocks
7 of the 99 YC startup homepages that enforce a Content Security Policy ask browsers for violation reports; 2 of those use report-to. The other 92 enforce a policy and never hear what it blocks, and 56 of those 92 would block at least one of the 11 analytics and replay tools as written (52 would block UserProof). Report-only policies are the exception: 6 of the 9 sites that send a Content-Security-Policy-Report-Only header collect its reports, as that header is meant to.
| YC homepages, read November 17, 2026 | Sites |
|---|---|
| Enforce a CSP (header or meta tag) | 99 |
| ... with report-uri or report-to | 7 |
| ... with report-to | 2 |
| ... with neither, and blocking at least one of the 11 tools | 56 |
| Send a report-only header | 9 |
| ... that collect its reports | 6 |
Source: UserProof, the home pages in our YC homepage benchmark read November 17, 2026 (data/yc-csp-2026-11-17.json: 376 fetched, 371 answered), counted by scripts/build-csp-report-reader.py (data/yc-csp-reporting-2026-11-18.json). Tools blocked: the CSP checker's rules on the header policy, else the meta tag.
What it can't do
- It doesn't collect reports. You need an endpoint that receives them; this page reads what it stored.
- A blocked URL on another origin arrives cut to scheme, host and port, so two products on the same host look the same. The reader names the tool whose CSP page lists that host.
- Inline code can't be traced to a tool without
'report-sample'; with a sample, the tool name is a guess from the text and says so. - It knows 11 tools' hosts. Other hosts get their origin as the source to add; check that you meant to load them before you allow them.
- The fixed policy never gains
'unsafe-inline'or'unsafe-eval'. It also never putsdata:orblob:inscript-srcordefault-src. Those weaken the policy, so the reader names them and leaves the call to you.
Where UserProof fits
UserProof records real visits with one external script tag and hands your coding agent the one fix to make, with the file and line. Under a CSP it needs two sources and nothing inline: https://ingest.getuserproof.com in script-src and connect-src. Field values are never recorded. Free covers 3 sites and 5,000 recorded visits a month; Pro is $29 a month for 50,000. To build a policy that lets your tools load, use the CSP generator; to test one, the CSP checker. All free tools are on the tools page.
Sources, read November 18, 2026
- MDN: report-uri: deprecated, the application/csp-report fields, cross-origin URLs cut to the origin. Also from that page: script-sample needs 'report-sample', and the directive doesn't work in a meta tag. MDN's report-to (Baseline since March 2026; csp-violation reports as application/reports+json; browsers that support it ignore report-uri) and Reporting-Endpoints (Baseline since September 2024).
- MDN: Content-Security-Policy (the fallback lists the reader walks).
- Vendor hosts: each vendor's own CSP page, listed with its date on the CSP generator's sources, read November 17, 2026.
- UserProof: 99 enforced policies from YC startup homepages read November 17, 2026, data/yc-csp-2026-11-17.json.
Cite this
UserProof, "CSP report reader," https://getuserproof.com/csp-report-reader; 7 of 99 enforced YC startup Content Security Policies send violation reports, read November 17, 2026.
FAQ
What is a CSP violation report?
A JSON message the browser sends when a page's Content-Security-Policy blocks something, or would block it under a report-only policy. It names the blocked URL (cut to the origin when it is another site's), the directive that blocked it, the page and the policy. You get them only if the policy has a report-uri or report-to directive.
How do I read a CSP report?
Look at three fields: blocked-uri (blockedURL in the Reporting API) says what was stopped, effective-directive says which rule stopped it, and disposition says whether it was blocked (enforce) or only reported (report). Paste the report above and the reader names the tool and the source to add.
What is the difference between report-uri and report-to?
report-uri takes a URL and posts application/csp-report JSON. report-to takes an endpoint name defined in the Reporting-Endpoints header and posts application/reports+json. MDN marks report-uri deprecated and says browsers that support report-to ignore report-uri, so send both until every browser you care about supports report-to.
Why do I get CSP reports from chrome-extension?
A browser extension on the visitor's computer injected a script or style into your page and your policy blocked it. It is not your code and there is nothing to add. Filter those reports out at your endpoint.
What does blocked-uri inline mean?
An inline <script> block, an inline event handler such as onclick, or an inline style was blocked. For scripts, put a nonce or a hash on the tag; 'unsafe-inline' works too but lets any injected script run. Add 'report-sample' to the directive to get the first 40 characters of the blocked code in the report.
Does the reader send my reports anywhere?
No. It runs in your browser, and the form is excluded from UserProof's own recorder on this page.
How many sites collect CSP reports?
Few. Of the 99 YC startup homepages we found enforcing a CSP on November 17, 2026, 7 have a report-uri or report-to directive. The other 92 enforce a policy and never hear what it blocks; 52 of those would block UserProof's script as written.