High-severity SSTI vulnerability in compliance-trestle's Jinja2 rendering pipeline

The oscal-compass/compliance-trestle project has disclosed a high-severity Server-Side Template Injection (SSTI) vulnerability, as detailed in the GitHub Security Advisory database. This issue arises from a systemic pattern in the project's Jinja2 rendering pipeline, where untrusted runtime data is re-parsed as Jinja2 template source code without adequate sandboxing or escaping of syntax delimiters. Specifically, the vulnerability allows injected expressions to traverse Python object chains and execute arbitrary commands via `os.system()` or `subprocess`.

Previously, a similar vulnerability in the `render_template()` function was addressed by limiting Jinja2 rendering to a single iteration. However, this advisory highlights two still-vulnerable code paths: `MDCleanInclude.parse()` and `MDSectionInclude.parse()`. Both functions involve re-evaluating untrusted Markdown content as Jinja2 source code without employing `SandboxedEnvironment` or escaping syntax delimiters, leaving them susceptible to SSTI attacks.

Operators should carefully assess their use of trestle, particularly if workflows involve rendering Markdown files or sections through these affected functions. Any data passed into these rendering pipelines should be scrutinized for potential injection risks. Before upgrading or applying patches, operators should test their existing templates and workflows to ensure compatibility with any changes to the rendering process. This vulnerability underscores the importance of avoiding recursive template re-evaluation patterns, a common issue in projects utilizing Jinja2.