Work at Anthropic? Claim it to respond to reviews as the verified owner.
Found a vulnerability?
If you would rather not deal with the vendor yourself, a BugRater analyst will submit it upstream on your behalf, with your explicit permission, and tell you what came back.
Ask BugRater to submit itPrivate. The report body is encrypted at rest; BugRater holds the key, so the analyst working it can read it. Every read is logged.
HackerOne’s own figures for this program, read from its public page, not reported by researchers and not part of the BugRater grade. Captured 22 Sep 2026.
What it pays, by severity
$1,059,485 paid to researchers in total, $292,500 of it in the last 90 days. Lifetime figure as HackerOne prints it: evidence this program has paid, not a promise about any one report.
Intake & responsiveness · last 90 days
Response targets it sets itself
A target the program declared, not a measurement of it being met.
Getting in the door
Open to submit. Nothing HackerOne publishes stands between a hunter and a first report here.
Over 31 days (17 snapshots): intake up 180 reports; response efficiency down 1 points; 90-day payout down $36,950.
See how this programme’s report load compares to others →
Reviews
0 publishedNo reviews yet.
Who this program credits
508 creditedResearchers HackerOne shows on this program’s public thanks list, best position first. “Recognised” is how many of a hunter’s submissions the program accepted; the ratio is their signal here, not our judgement of them.
| # | Researcher | Reputation | Recognised / submitted |
|---|---|---|---|
| 1 | ryotak | 2,817 | 62 / 92 67% |
| 2 | stopwar | 659 | 16 / 32 50% |
| 3 | liamconner10 | 642 | 4 / 17 24% |
| 4 | eran667 | 531 | 10 / 13 77% |
| 5 | amr_id | 454 | 15 / 44 34% |
| 6 | edbr | 345 | 8 / 10 80% |
| 7 | lucasfutures | 308 | 13 / 75 17% |
| 8 | mahoraga007 | 291 | 3 / 47 6% |
| 9 | realmarmarlabs | 278 | 11 / 45 24% |
| 10 | norwegianlamb | 275 | 10 / 130 8% |
| 11 | orenyom | 270 | 6 / 7 86% |
| 12 | anagnorisis2peripeteia | 256 | 1 / 86 1% |
| 13 | philts | 244 | 7 / 8 88% |
| 14 | cotu223 | 236 | 6 / 62 10% |
| 14 | fabiusartrel | 236 | 4 / 5 80% |
| 16 | argareksapatii | 233 | 7 / 27 26% |
| 17 | waynezinn | 232 | 4 / 30 13% |
| 18 | yottt | 214 | 8 / 43 19% |
| 19 | n9j3h9u8 | 196 | 8 / 14 57% |
| 20 | m0chan | 190 | 8 / 11 73% |
| 21 | wunderwuzzi23 | 184 | 4 / 10 40% |
| 22 | suul | 183 | 10 / 34 29% |
| 23 | lachlan2k | 175 | 3 / 6 50% |
| 24 | cantina-security | 174 | 5 / 40 13% |
| 25 | d0xing | 173 | 6 / 8 75% |
Showing the top 25 of 508 credited on HackerOne.
Facts published by HackerOne on the program's own page, not reported by researchers, and not part of the BugRater grade. Last checked 16 Sep 2026.
Anthropic is an AI safety and research company that builds reliable, interpretable, and steerable AI systems.
Scope
22 assets| Asset | Type | Eligibility | Max severity |
|---|---|---|---|
| anthropic.atlassian.com | URL | ✓ bounty | Critical |
| API & SDKs | AI MODEL | ✓ bounty | Critical |
| api.anthropic.com | URL | ✓ bounty | Critical |
| Claude iOS app | APPLE STORE APP ID | ✓ bounty | Critical |
| Claude Code | OTHER | ✓ bounty | Critical |
Show all 22 assets
| Asset | Type | Eligibility | Max severity |
|---|---|---|---|
| Claude Desktop Extensions and Claude.ai MCP servers | OTHER | ✓ bounty | Critical |
| Claude in Chrome | OTHER | ✓ bounty | Critical |
| Claude Model Safety | OTHER | ✓ bounty | Critical |
| claude.ai | URL | ✓ bounty | Critical |
| console.anthropic.com | URL | ✓ bounty | Critical |
| Core Assets | OTHER | ✓ bounty | Critical |
| Core Assets | | OTHER | ✓ bounty | Critical |
| docs.anthropic.com | URL | ✓ bounty | Critical |
| github.com/anthropics | SOURCE CODE | ✓ bounty | Critical |
| https://apps.apple.com/us/app/claude/id6473753684 | APPLE STORE APP ID | ✓ bounty | Critical |
| Infrastructure & Internal Apps/Services | OTHER | ✓ bounty | Critical |
| Leaked Employee API Keys | OTHER | ✓ bounty | Critical |
| Non-Core Assets | OTHER | ✓ bounty | Critical |
| Official Clients | OTHER | ✓ bounty | Critical |
| support.anthropic.com | URL | ✓ bounty | Critical |
| github.com/modelcontextprotocol/servers-archived | SOURCE CODE | out | None |
| https://github.com/modelcontextprotocol | SOURCE CODE | out | None |
HackerOne lists 23 scope entries; its public listing groups many assets under one label, so identical entries are shown once.