- Top reward
- $20,000
- Top severity paid
- Critical
- Assets in scope
- 23
- Public CVEs
- 0
- Last change
At a glance
What to know before you start testing.
- Pays up to $20,000
- for critical bugs.
- Report on HackerOne
- Submit through the program page.
- Replies within 2 days
- As the policy promises.
- No public write-ups
- Keep findings private.
Rewards
What HackerOne lists as the payout for each severity.
| Severity | Relative size | Reward |
|---|---|---|
| Critical |
|
$20,000 from $7,500 |
| High |
|
$7,000 from $2,500 |
| Medium |
|
$2,000 from $500 |
| Low |
|
$500 from $100 |
- Response
-
Within 2 days
First Response: 2 business days
Summarised with AI from the program policy on HackerOne, read 12 Nov 2026, and checked word for word against it. Read the original
Scope
23 assets you may test, 1 excluded.
Web 12
- accounts.x.ai Critical
- chat.x.com Critical
- console.x.ai Critical
- console.x.com Critical
- grok.com Critical
- grokipedia.com Critical
- money.x.com Critical
- pscp.tv Critical
- t.co Medium
- xadsacademy.com Medium No bounty
- x.ai Critical
- x.com Critical
Wildcard domains 6
- *.grok.com Critical
- *.twimg.com Critical
- *.twitter.com Critical
- *.vine.co Critical
- *.x.ai Critical
- *.x.com Critical
Android apps 2
- ai.x.grok Critical
- com.twitter.android Critical
iOS apps 2
- ai.x.GrokApp Critical
- com.atebits.Tweetie2 Critical
Products and areas 1
- X Pixel Helper (Chrome Extension) Critical
Out of scope (1)
- grok-build
Rules
What the policy asks of you, and the bug types it won't reward, quoted from X / xAI (via HackerOne)'s policy. "We" in these lines means the program, not this site.
While testing
-
you must use test accounts in order to respect our users’ privacy.
-
You must be the first reporter of the vulnerability.
-
You must not have publicly disclosed the vulnerability prior to the report being closed, in compliance with the process described in the HackerOne Vulnerability Disclosure Guidelines.
-
Subdomain takeovers - please demonstrate that you are able to take over the page by leaving a non-offensive message, such as your username.
Not accepted
-
Highly speculative reports about theoretical damage.
-
Vulnerabilities as reported by automated tools that have not been validated.
-
SSL/TLS scan reports (ie: output from sites such as SSL Labs).
-
Banner grabbing issues (figuring out what web server we use, etc.)
-
Volumetric Denial of Service (DoS) attacks.
-
Forms missing CSRF tokens (we require evidence of actual CSRF vulnerability).
-
Logout CSRF.
-
Password and account recovery policies, such as reset link expiration or password complexity.
-
Invalid or missing SPF (Sender Policy Framework) records.
-
Bypass of URL malware detection.
-
Open redirects unless they can demonstrate a higher security risk than phishing.
-
Account oracles -- the ability to submit a phone number, email, UUID and receive back a message indicating an X/xAI account exists.
-
Reports that state that software is out of date/vulnerable without a proof-of-concept.
-
Model outputs that violate content moderation guardrails.
-
Performing code execution on grok.com in the sandboxed environment.
-
CSRF with linked queries (grok.com?q=prompt) that only create a new grok chat.
-
Grok Imagine media is publicly accessible to anyone with the link, and reports of that access are ineligible.
-
Issues focusing on the client-side bypass of auto-approval for safe/trusted commands are ineligible for bounty.
-
Manipulation of "likes/follows/views" count due to caching issues.
-
Homoglyph attacks in URLs that can lead to phishing scenarios, unless the finding can demonstrate larger issues on our platform (e.g AuthN/AuthZ bypass)
Summarised with AI from the program policy on HackerOne, read 12 Nov 2026, and checked word for word against it. Read the original
Changes
Tracked since 11 Nov 2026. Last checked 1 hour ago.
No changes yet. When the rewards or scope move, you will see it here and in the RSS feed.
Before we tracked it
Scope and reward changes recorded before this page started tracking it.
-
Scope removed
grok-build-cli
-
Scope removed
gnip.com
-
Scope removed
status.twitter.com (out of scope)
-
Scope added
accounts.x.ai
-
Scope added
console.x.ai
-
Scope added
pscp.tv
-
Scope added
X Pixel Helper (Chrome Extension)
-
Scope added
x.ai
-
Scope removed
*.twitter.biz
-
Scope added
console.x.com
Public CVEs
Published vulnerabilities in this vendor's products, from the NVD.
No CVEs matched yet. We match NVD records by vendor name, so products sold under another name can be missing.
Original policy
Copied from hackerone.com on 12 Nov 2026. The summaries on this page are made from it; the live page may have changed since.
Program Rules
The scope of this program is limited to products under under X and xAI. Please make sure you review the following program rules before you report a vulnerability. By participating in this program, you agree to be bound by these rules.
Rewards
- SpaceXAI may, at its sole discretion, provide rewards to eligible reporters of qualifying vulnerabilities.
- Issue severity is calculated by an internal risk matrix based on company specific data and use cases. Severity is considered a combination of Impact and Likelihood, each assigned a value of Informative, Low, Medium, High, or Critical.
- Payouts are determined by a panel of security experts. Previous bounty rewards are not considered as precedent for future rewards.
Legal Fine Print
- You must comply with all applicable laws in connection with your participation in this program. You are also responsible for any applicable taxes associated with any reward you receive.
- We may modify the terms of this program or terminate this program at any time. We won’t apply any changes we make to these program terms retroactively.
SLA
SpaceXAI will make a best effort to abide by the following timelines: First Response: 2 business days Time to Triage: 10 business days Time to Bounty: 14 business days Time to Resolution: depends on severity and complexity
Rules of Engagement
When researching security issues, especially those which may compromise the privacy of others, you must use test accounts in order to respect our users’ privacy. Accessing private information of other users, performing actions that may negatively affect SpaceXAI users (e.g., spam, denial of service), or sending reports from automated tools without verifying them will immediately disqualify the report and may result in additional steps being taken.
Report Eligibility
General
- You must be the first reporter of the vulnerability.
- The vulnerability must demonstrate security impact to a site or application in scope.
- You must not have compromised the privacy of our users or otherwise violated the SpaceXAI Rules.
- You must not have publicly disclosed the vulnerability prior to the report being closed, in compliance with the process described in the HackerOne Vulnerability Disclosure Guidelines.
- We must not be legally prohibited from rewarding you.
Open-Source Code
Reports for the open-source SpaceXAI code have additional requirements. A working proof of concept against our services is required for report acceptance. If a proof of concept is not provided, nor sufficient supporting documentation for SpaceXAI engineers to feasibly recreate and evaluate the issue, the report may not be accepted. Report quality for acceptance will be at the discretion of SpaceXAI's HackerOne program. If you can not showcase an active exploited against any of our hosted services, and if you believe that the Open Source repository is actively maintained, please submit a PR on the repository.
Ineligible/Known Issues
- Attacks requiring physical access to a user's device.
- Any physical attacks against SpaceXAI property or data centers.
- Social engineering of SpaceXAI staff or contractors.
- Highly speculative reports about theoretical damage.
- Vulnerabilities as reported by automated tools that have not been validated.
- SSL/TLS scan reports (ie: output from sites such as SSL Labs).
- Banner grabbing issues (figuring out what web server we use, etc.)
- Volumetric Denial of Service (DoS) attacks.
- Issues without clearly identified security impact, such as clickjacking on a static website, missing security headers, or descriptive error messages.
- Forms missing CSRF tokens (we require evidence of actual CSRF vulnerability).
- Issues related to software or protocols not under SpaceXAI control.
- Logout CSRF.
- Vulnerabilities only affecting users of outdated or unpatched browsers and platforms.
- Reports of broken hyperlinks from X blog posts, Grokipedia pages, press releases, or support articles to unclaimed X Handles.
- Password and account recovery policies, such as reset link expiration or password complexity.
- Invalid or missing SPF (Sender Policy Framework) records.
- Bypass of URL malware detection.
- Issues relating to unlocking client-side features in modified SpaceXAI applications, rooted devices, or jailbroken devices.
- Open redirects unless they can demonstrate a higher security risk than phishing.
- Account oracles -- the ability to submit a phone number, email, UUID and receive back a message indicating an X/xAI account exists.
- Reports that state that software is out of date/vulnerable without a proof-of-concept.
Grok
- Model outputs that violate content moderation guardrails. These should be reported through [email protected]
- Performing code execution on grok.com in the sandboxed environment.
- CSRF with linked queries (grok.com?q=prompt) that only create a new grok chat.
- Grok Imagine media is publicly accessible to anyone with the link, and reports of that access are ineligible.
Grok Build
- Issues focusing on the client-side bypass of auto-approval for safe/trusted commands are ineligible for bounty.
X
- Manipulation of "likes/follows/views" count due to caching issues. We are already aware that some statistics on our site may temporarily display inaccurate figures when many requests are sent in a brief period of time. However, that is because the numbers are cached, and the cache may take a bit of time to synchronize with the accurate backend data store.
- Homoglyph attacks in URLs that can lead to phishing scenarios, unless the finding can demonstrate larger issues on our platform (e.g AuthN/AuthZ bypass)
- Documents in the Ads Transparency Center do not disclose private account information. We intentionally disclose the advertising status of accounts to better inform our users of ads being run our platform.
Eligible Issues
Some of the aforementioned guidelines may put your report in a gray area. Reports demonstrating the following might be considered valid :
- Subdomain takeovers - please demonstrate that you are able to take over the page by leaving a non-offensive message, such as your username.
- We only consider DoS issues that can be triggered by a single user with a single or few requests.
- Valid employee credentials from data dumps.
Grok
- Model interactions that cross the security boundary of chat responses and somehow affect Confidentiality, Integrity, Availability of other users or SpaceXAI.
- CSRF with linked queries (grok.com?q=prompt) that can interact with user's account outside the single chat, interact with settings, connectors or exfiltrate memory.
Grok Build
- Bypassing security controls baked into Grok Build binary - like escaping sandbox profiles.
- Reports demonstrating command executions at Grok Build startup bypassing Trust gate may be considered for a bonus.
Report Disclosures
We currently don't disclose reports submitted to our program.