Where the deeplink/nonce work can actually make money, and the one reporting trap that gets this entire bug class closed as N/A. Desk research only: public scope pages and disclosed reports, no testing of any of these targets.
The pattern worth carrying forward
The deeplink trace left me with one reusable
idea: a server-rendered value interpolated raw into a single-quoted string
inside a nonce'd inline <script>. The CSP doesn't stop it, because the
script was already trusted. The bug isn't "weak CSP": it's a trusted script
doing the injection for you.
That shape is portable. The question is which programs pay for it, and, more importantly, which programs pay for it without rejecting it as a header complaint.
Basecamp / HEY is the closest fit
On HackerOne, Basecamp's program lists injection prevention (SQL, XSS, method args) as a focus area, pays $100–$10,000, and explicitly says "concatenating bugs to increase the attack scenario is encouraged." That last line is an invitation to the chain-style work this class needs.
Their disclosed history proves they mean it:
- 2020,
app.hey.comsearch box: a DOM XSS where CSP blocked execution; the report's whole point was if the CSP can be bypassed, it fires. Triaged, paid, bumped to high. Staff wrote: "We are interested in dealing with root causes for potential XSS issues, even when intercepted by our CSP." - 2020, stored XSS in hey.com: a sanitizer bypass mitigated by CSP. They reversed policy to treat sanitizer bypasses as eligible separately from CSP bypasses, because the two defenses are graded on their own.
- 2026, cross-account XSS on
app.basecamp.com(CVSS 8.7, disclosed 2026-10-28): an oEmbed option injected unescaped, then an allowlisted Basecamp script read the page'smeta[name=csp-nonce]and stamped it onto the attacker's script. The nonce was never guessed. As the write-up put it, a nonce is only as secret as the DOM that carries it.
That 2026 report is the closest public analogue to our pattern: trusted script, nonce laundering, no CSP break. If the deeplink bug existed on their origin, it would be squarely in their wheelhouse.
Intigriti's public list, honestly
Checking the live program list on 2026-11-10, the paying options are thin:
- Lightspark: up to $5,000, payments infrastructure.
- Adobe: $75–$15,000, 2FA required to participate.
Everything else public (Coke One North America, Rivian, BorgWarner, NxtPort, Intel, ADAC, Dutch Lottery, Mateco) is a VDP with no bounty. Good for reputation and write-ups, zero payout. Worth knowing before you spend a session.
Intigriti's real value for this class is its monthly challenges and Bug Bytes research, which are a free, explicitly authorized syllabus for nonce-bypass and DOM-clobbering chains. Rehearse there; get paid on Basecamp.
The trap: "CSP headers" is out of scope, "XSS despite CSP" is not
This is the single most important distinction in the whole effort.
Basecamp's out-of-scope list includes "HSTS or CSP headers." Read that carefully. It does not mean CSP-bypass XSS is out of scope. It means:
- ❌ Filing "your CSP is missing
object-src/ allowsunsafe-inline" → closed as N/A, every time. - ✅ Filing "here is script execution the CSP did not stop, with impact" → in scope, and paid.
The header is not the finding. The execution is. Any report that leads with "the CSP is weak" has already lost; the CSP's job is to catch bugs, and a bug that it correctly catches (or that executes inside a trusted nonce'd script) is only reportable if you show the impact regardless of the header.
I've written a reusable report template shape for this class in the engagement notes: source, sink, why the nonce makes CSP irrelevant, deterministic repro with two accounts you own, and a scope checklist that refuses to file on a header complaint.
What "making cash" actually requires
Not volume. All three programs above reward impact and understanding and punish noise. A realistic plan:
- Read the live Basecamp scope page end to end; confirm it's open to me.
- Create two throwaway accounts (attacker + victim) with the email convention they ask for.
- Run the nonce-laundering checklist against the in-scope assets, minimal requests, everything logged.
- Only file if a quote actually breaks out of a nonce'd script with impact.
One well-documented nonce-laundering chain beats fifty "missing header" reports, and it's the only version of this work that pays.