If you find a security vulnerability in Cookiecutter, please report it through GitHub's private vulnerability reporting. This keeps the details private while we work on a fix.

Please include:

• What you found and how to reproduce it
• Which version you're using
• Any relevant logs or output (redact secrets)

Cookiecutter executes templates from sources the user specifies (GitHub repos, zip URLs). Templates can run arbitrary code through hook scripts (pre_prompt, pre_gen_project, post_gen_project), and Cookiecutter does not sandbox them. This is by design. Users are responsible for reviewing templates before running them.

• The cookiecutter CLI and Python API
• Jinja2 template rendering and variable handling
• Hook execution (pre_gen_project, post_gen_project, pre_prompt)
• Config file handling (~/.cookiecutterrc, cookiecutter.json)
• Git and zip-based template retrieval
• Hook execution awareness bypass (mechanisms that cause hooks to run without the user's knowledge)
• Template rendering exfiltration (vulnerabilities in the Jinja2 layer that leak data outside the expected output)

If you're unsure whether something qualifies, report it and we'll assess.

• Hook scripts executing arbitrary code (this is intended behavior, see Trust Model above)
• Malicious content in third-party templates (template authors are outside our trust boundary)
• Vulnerabilities in dependencies of generated projects (those belong to the template author)

• Dependabot monitors GitHub Actions dependencies for known vulnerabilities
• Least-privilege CI permissions (permissions: contents: read at the workflow level)
• Hook validation only recognizes known hook names (pre_prompt, pre_gen_project, post_gen_project), rejecting unexpected scripts in the hooks/ directory
• Failed hook cleanup automatically removes partially-generated project directories when a hook script fails, preventing incomplete output from being used

Cookiecutter does not collect, store, or transmit any user data. There is no telemetry, analytics, or phone-home behavior of any kind.

This is a volunteer-maintained open source project. Security reports are taken seriously, but there are no guaranteed response times.

Enterprise support is available, with priority response SLAs. Contact support@feldroy.com for details.

Security fixes are applied to the latest release on the main branch. There is no backport policy for older versions.