FAQ
Get ready to research.
Answers to common questions about the program, your report, and what comes next.
How is this different from a penetration test?
They work together. A pen test is a scoped, time-boxed vendor engagement that produces a formal report. Avala runs those annually, separately from the bounty program. The bounty program pays for eligible findings. New researcher intake is currently closed; existing researchers can continue using the portal. Bounty findings feed back into the next pen test scope.
Do I need an account to submit?
Yes. We use magic-link sign-in (no password) so every report has a private thread, status tracking, and a payout channel. We are not accepting new researchers at this time. Existing researchers can still sign in. For urgent vulnerability disclosures, email security@avala.ai. We store your email, display name, and (once a bounty is approved) payment details.
Can I remain anonymous?
Yes. On each report you can choose to be credited by real name, pseudonym, or anonymously. If you want to receive a cash bounty we'll need a PayPal address and a W-9 or W-8BEN, but your public attribution on hall-of-fame / CVE credits can still be anonymous.
What CVSS version do you use?
CVSS v3.1. The severity table in Rewards maps CVSS bands to cash ranges. We accept your self-assessment as a starting point and may adjust the score during triage.
Do duplicates get paid?
First-to-report wins the bounty. Duplicates receive CVE credit if applicable and a thank-you, but not cash. If you believe your report provides materially different information or a better exploit chain, flag it in the submission and we'll review.
How long until I get paid?
There is no fixed payment timeline. Approved bounties are recorded and paid in full, in the order reports were received, as the monthly pool allows. Payouts are currently delayed, and the program status notice carries the latest update. All payouts are via PayPal, with no fees deducted from the awarded amount. A W-9 (US) or W-8BEN (non-US) form is required before first payout.
How quickly will you triage or fix my report?
We do not publish response-time targets, and we are a small team working through reports in the order they arrive. Every report is acknowledged, every validated report is fixed, and every approved bounty is paid. What we cannot do is promise when. If your report has been quiet, it is still open; you are welcome to ask for a status in the report thread.
Can I use AI tools or an AI agent in my research?
Yes. AI-assisted research is eligible when a human has independently reproduced the finding against the published asset or package and the report says so — include the exact version tested, the reproduction steps you ran yourself, and the evidence. Reports generated primarily by an automated tool or LLM without that independent validation are not eligible, and an AI agent cannot hold a researcher account or receive a payout on its own; the account, tax forms and payment belong to a person.
Can I publicly disclose my finding?
Yes, once the fix is deployed and we have agreed a publication date in your report thread. Until then the finding stays confidential, including from third parties. If a validated report has been open for 90 days without a fix, ask for a date in the thread and we will propose one. Publishing without an agreed date forfeits payment on that report and on your other reports, and ends your participation in the program. Full terms are in the rules under disclosure and payment eligibility.
Can the program terms change after I submit?
Yes. Avala may change the program at any time, and the version on the rules page with its effective date is the one in force. Reward tiers for an open report stay at the tiers published when you submitted it. The eligibility and payment-process terms in force when we make the bounty decision govern that decision. Open reports keep the disclosure obligation you agreed to at submission.
What about AI/LLM vulnerabilities?
These are explicitly in scope. Indirect prompt injection via customer-uploaded annotation data, authz bypass in our MCP server, model exfiltration through prompts, and novel AI-specific attack classes are particularly valuable to us and eligible for the novel-class multiplier.
Can you sponsor my research with grants or equity?
Right now: cash bounties and public recognition. Researchers with 3+ valid criticals may be considered for a paid security advisor engagement. If that sounds interesting, reach out after your first few reports land.
I have a question before I submit.
Email security@avala.ai with your pre-submission question. For anything that involves a working exploit, please submit through the portal to keep it in a secure thread.
Contact
Need a specific answer?
Email the security team with program questions. Submit findings through the portal so evidence and follow-up stay in a private thread.