Cyberpediaby Brain

Responsible disclosure and bug bounty

Report a flaw privately, give time to fix it, then publish: how disclosure policies and bounty programs work, and what it means for you.

On this page
  1. 1In 30 seconds
  2. 2What it is
  3. 3Straight answers
  4. 4What it changes for employees
  5. 5What it changes for management
  6. 6What the text says
  7. 7Where implementation stands
  8. 8Sources
  9. 9Cite this article

In 30 seconds

Where it happens
You may find a flaw as part of your work, or receive a report from an outside person; staff who deal with the public must know who to pass such a report to[1].
The reflex
Stop the test and notify the organization concerned right away, using the contact given in its policy; do not share any data you came across[4].
If it's already done
If you stumbled on sensitive data, stop, notify immediately and do not share it with anyone else[4].

What it is

Responsible disclosure means reporting a flaw privately to the organization concerned, then publishing the details once a fix exists, sometimes with an extra delay to give people time to install it[1]. It sits between two extremes: private disclosure, where the organization alone decides whether to publish, and full disclosure, where everything is made public as soon as the flaw is found[1].

The debate is old. On 20 July 2010, Google published an article on its security blog titled "Rebooting Responsible Disclosure: a focus on protecting end users"[2]. Microsoft's bounty program was officially launched on 26 June 2013, and in November 2023 the company announced more than 60 million dollars paid to researchers since 2013[3].

Straight answers

What is responsible disclosure?

The initial report is made privately. The full details are published only once the fix is available[1]. Often the researcher also sets a deadline for the organization to respond or provide a fix; if it is not met, the researcher may move to full disclosure[1]. The Google Project Zero policy, cited by OWASP, generally gives the vendor 90 days to provide a fix, then publishes the technical details 30 days after that fix; if there is no fix by the deadline, publication normally takes place at 90 days[1].

How does it differ from private and full disclosure?

In the private model, the organization can choose never to publish anything; most bounty programs require this model[1]. If the vendor does not respond or gives up on fixing the flaw, it may stay ignored, and this is what pushed researchers toward full disclosure[1]. Full disclosure exposes the details, sometimes the exploit code, to attackers, often before any fix. Many consider it irresponsible, and according to OWASP it is justified only as a last resort or if exploit code is already public[1].

What is coordinated disclosure?

According to Microsoft, the researcher reports the flaw to the vendor, to a national CERT or to another coordinator who will notify the vendor privately. The vendor has time to diagnose the problem and provide fixes or workarounds before anyone publishes details[5]. If attacks are already under way while the fix is not ready, the researcher and the vendor work together on an early public disclosure[5]. CISA states for its part that when a vendor does not respond, it may disclose a flaw as early as 45 days after the first contact attempt, even without a fix[6].

What is a bug bounty?

A bounty program offers rewards to researchers who report flaws. They are most often financial, sometimes in goods[1]. A third party such as BugCrowd or HackerOne often acts as an intermediary between researchers and the organization[1]. At Microsoft, bounties go up to 250,000 dollars for the internal workstation and server programs, and up to 100,000 dollars for the cloud ones[7].

What is a disclosure policy for?

According to CISA, it sets how an organization receives reports, what is in scope and what the reporter can expect[6]. In CISA's model, the organization commits not to recommend or pursue legal action for research conducted in good faith under the policy[4]. It also promises to acknowledge receipt within 3 business days, if the reporter leaves contact details[4].

What it changes for employees

If you discover a flaw:

  • Do not test beyond what is necessary. CISA's model asks you to use an exploit only to the extent needed to confirm the flaw, without extracting data or pivoting to other systems[4]. It also rules out denial-of-service testing, physical testing and social engineering[4].
  • Stop when you see sensitive data. Once the flaw is established, or as soon as you encounter sensitive data, you stop the test and notify without delay[4]. Microsoft gives the same instruction: if in doubt about what is safe, you stop and contact them[7].
  • Do not publish. CISA asks for a reasonable delay before any public disclosure[4]. Many bounty programs forbid publication without the organization's agreement, at the risk of losing the bounty or being excluded[1].
  • Pass it quickly to the right contact. Avoid handing the details to the first person who answers, and ask instead for the name of a competent colleague[1].
  • Read your contract. If you find the flaw as part of your job or on your employer's equipment, your employer may forbid you from reporting it or claiming a bounty[1].
  • Stay within scope. Going beyond the scope and rules of a program can be a criminal offense[1].

If you receive a report, pass it to the designated internal contact and keep the reporter informed of progress[1]. If someone demands payment before handing over the details, it may be a scam; OWASP suggests asking them to go through a bounty platform with mediation[1]. Demanding payment in exchange for details or for silence can also be treated as extortion[1].

What it changes for managementManagement

OWASP asks organizations to offer a clear and safe way to report flaws, to respond within a reasonable time, to communicate openly, not to threaten researchers with legal action and to publish clear security advisories[1]. The most important step is to publish a way to make contact: a security@ address, a dedicated contact page, a security.txt file or a third-party bounty program[1]. Front-line staff (general mailbox, chat, phone) must know how to handle a report and who to escalate it to[1].

CISA's model specifies what a policy contains: an authorization section free of legal jargon, a scope, the excluded testing methods, the information expected in a report and the organization's commitments[4]. It requires that reporting can be done anonymously, without demanding personal data[4]. Before including a service managed by a provider in scope, the organization must make sure the provider allows testing; otherwise the service stays excluded[4].

A bounty program asks for more. OWASP reserves it for organizations whose disclosure process is already mature: it takes time to respond, staff able to triage, and it brings in many false positives[1]. The scope, the types of flaws that earn a reward, the legal safeguards and the amounts must be defined in it[1]. As for publishing an advisory after a fix, OWASP writes that it does not make the vendor look bad: a known handling process builds more trust than silence[1].

What the text saysOfficial text

The frameworks come from several origins.

  • United States, CISA. The model policy is that of Binding Operational Directive BOD 20-01: it must be published as a public web page at the path "/vulnerability-disclosure-policy" of the agency's main .gov site, and address all the actions required by the directive[4]. CISA also runs a coordinated disclosure program, separate from the disclosure policy: the former coordinates flaws, fixes and advisories between researchers and vendors; the latter only defines how an organization receives reports about its own systems[6].
  • United States, NIST. Publication SP 800-216 recommends a federal disclosure framework, the handling of reports and communication about fixes; it is intended to apply to all software, hardware and digital services under federal control[8].
  • Standards. Microsoft points to the ISO/IEC 29147:2018 standard on vulnerability disclosure[5].
  • European Union, Cyber Resilience Act. According to ENISA, the obligations of Article 14 cover actively exploited vulnerabilities and severe incidents. A manufacturer reports them once, through a single platform, instead of notifying several national authorities[9].

Where implementation stands

  • Cyber Resilience Act: the reporting obligations of Article 14 have applied since 11 September 2026[9]. The single reporting platform is a tool established by ENISA[9].
  • ENISA: it has been a CVE Numbering Authority (CNA) since January 2024 and a root of the CVE program since 20 November 2025[9]. It has published guides to help Member States set up coordinated disclosure policies[9].
  • NIST: SP 800-216 was published as a draft on 7 June 2021, then in final version on 24 May 2023[8].
  • CISA: directive BOD 20-01 provides that 2 years after its publication, all internet-accessible systems of an agency must be within the scope of its policy[4].
  • Microsoft: every report that is triaged, assessed and fixed is reviewed for a possible bounty, with no prior enrollment[3].

Sources

  1. OWASP, “Vulnerability Disclosure Cheat Sheet”. https://cheatsheetseries.owasp.org/cheatsheets/Vulnerability_Disclosure_Cheat_Sheet.html (accessed 14 November 2026)
  2. Google, “Rebooting Responsible Disclosure: a focus on protecting end users”, 20 July 2010. https://security.googleblog.com/2010/07/rebooting-responsible-disclosure-focus.html (accessed 14 November 2026)
  3. Microsoft, “Celebrating ten years of the Microsoft Bug Bounty program and more than $60M awarded”, 20 November 2023. https://www.microsoft.com/en-us/msrc/blog/2023/11/celebrating-ten-years-of-the-microsoft-bug-bounty-program-and-more-than-60m-awarded (accessed 14 November 2026)
  4. CISA, “Vulnerability Disclosure Policy Template”. https://www.cisa.gov/resources-tools/resources/vulnerability-disclosure-policy-template (accessed 14 November 2026)
  5. Microsoft, “Coordinated Vulnerability Disclosure”. https://www.microsoft.com/en-us/msrc/cvd (accessed 14 November 2026)
  6. CISA, “The Coordinated Vulnerability Disclosure (CVD) Program”. https://www.cisa.gov/resources-tools/programs/coordinated-vulnerability-disclosure-cvd-program (accessed 14 November 2026)
  7. Microsoft, “Microsoft Bounty Programs”. https://www.microsoft.com/en-us/msrc/bounty (accessed 14 November 2026)
  8. NIST, “Recommendations for Federal Vulnerability Disclosure Guidelines”, 24 May 2023. https://csrc.nist.gov/pubs/sp/800/216/final (accessed 14 November 2026)
  9. ENISA, “Vulnerability Services”. https://www.enisa.europa.eu/topics/vulnerability-services (accessed 14 November 2026)