GitHub now offers scheduled dependency updates for agent skills and instructions. CodeQL 2.27.3 adds a way to model AI-generated data in static analysis, and both CodeQL and Dependabot add Python 3.15 support. These changes give teams more places to review what an upgrade changes before rolling it out.
Research window: three GitHub announcements dated 3 November 2026, from the preceding calendar day. Exact publication times were unavailable. Coverage stayed within that date.
Dependabot can open version-update pull requests for supported Git dependencies declared in apm.yml and regenerate apm.lock.yaml when present. Configure the apm ecosystem with the manifest directory and a schedule. GitHub.com supports this across all plans, and GitHub lists Enterprise Server 3.24 as the target. APM support covers version updates only. Review changes to prompts, skills and instructions in the pull request because they can change agent behavior.
https://github.blog/changelog/2026-11-03-dependabot-version-updates-support-agent-package-manager/
CodeQL 2.27.3 supports JSON data extensions and adds an opt-in ai threat model for LLM-generated data. Teams must model their own sources for that threat model. The release also changes the control flow graph used for GitHub Actions analysis, so custom queries tied to its structure may need updating. Validate custom source models and queries on examples with known expected findings.
https://github.blog/changelog/2026-11-03-codeql-2-27-3-adds-python-3-15-features-and-json-data-extensions/
Dependabot now supports Python 3.15 on GitHub.com, with Enterprise Server 3.24 planned. CodeQL also supports new Python 3.15 syntax and builtins. Before changing the runtime, check dependency resolution, application tests and scanning in the same upgrade branch. GitHub's tool support does not establish compatibility for any particular application.
https://github.blog/changelog/2026-11-03-dependabot-now-supports-python-3-15/