GitHub has made custom permissions for repository security advisories generally available, allowing organisation owners to assign security-related work without making every participant a repository administrator. The capability covers tasks involving security advisories, including private vulnerability reports. GitHub says the change aims to let teams delegate this work through custom repository roles rather than granting broader access than a contributor needs. The update is relevant to open-source maintainers and enterprise development teams that coordinate vulnerability handling across engineers, security specialists and external researchers.
Security advisory management often involves people who need access to sensitive information but do not need control over the entire repository. A security researcher may need to review a private report, a maintainer may need to coordinate a fix and a security lead may need to prepare an advisory for publication. If the only available permission is full administrator access, organisations may grant more authority than necessary. The GitHub announcement says custom roles can provide a narrower route for these tasks. The original reporting on GitHub sets out the primary development.
The change narrows a common privilege trade-off
The change reflects a principle of least privilege: each person should receive only the permissions required for their responsibilities. Broad repository access can expose source code, workflow settings, secrets-management controls and other administrative functions. A person who only needs to review a vulnerability report should not automatically receive the ability to change unrelated repository settings. More granular roles can reduce the consequences of an account compromise and make access reviews easier to understand.
Private vulnerability reports require particular care because they may contain reproduction steps, affected code paths, exploit details or information about systems that have not yet been patched. Access needs to be limited to the people involved in triage and remediation. Organisations also need a process for moving from private coordination to public disclosure when a fix is ready. Permission controls support that workflow, but they do not replace a disclosure policy or the judgement required to assess whether a vulnerability is safe to publish. Further technical context for Security advisories is available in the supporting documentation.
Maintainers should review existing repository roles and identify where administrators have been assigned primarily to handle security advisories. Those accounts may be candidates for a narrower custom role. Before changing permissions, teams should test whether the new role supports the full workflow, including viewing reports, commenting, assigning work and preparing advisory information. A role that is too restrictive can delay response, while one that is too broad can recreate the original problem under a different name. A runtime-security example appears in Cloudflare Adds WAF Protection for Ruby2Shell as Rails and Sinatra Users Face a Critical Memory Bug, which examines a separate vulnerability and edge mitigation.
The disclosure process needs a separate permission path
The distinction between advisory permissions and general repository permissions is important during a coordinated disclosure. A private report may need to be shared with a small group of maintainers before a patch is ready, while a public advisory may later be visible to a much wider audience.
Security work can be delegated without broad admin access
Organisations should define when access is expanded, who approves publication and how sensitive details are removed from the public record. Granular roles can support that separation, but the project still needs a clear disclosure process and a named person responsible for each stage.
The feature may be useful for projects with external security researchers or distributed maintainer teams. A project can involve volunteers, corporate engineers and researchers who contribute at different stages of a disclosure. A custom permission model can make it easier to grant access for a defined task without changing the person’s general repository status. Organisations should still use identity verification, multifactor authentication and audit logs because the sensitivity of private vulnerability reports makes account security essential. The source coverage of Repository permissions provides the underlying detail.
Security teams should also document who is responsible for reviewing reports, who can approve a fix and who decides when a public advisory is released. The permission model can enforce some boundaries, but the workflow needs clear ownership. If a report sits in a private queue without an assigned maintainer, granular permissions alone will not make the response faster. Teams should combine access control with response-time targets and a process for escalating reports that affect widely deployed software.
GitHub’s update is a focused change, but it addresses a real tension in software security: the people who need to coordinate vulnerability handling are not always repository administrators. More precise roles can reduce unnecessary privilege and support safer collaboration. The next step for maintainers is to review current access, test the new permissions and ensure that private reports remain visible only to the people who need them. Security advisory work should be easier to delegate without expanding the administrative attack surface.