prepare-code-freeze

Pass

Audited by Gen Agent Trust Hub on Oct 3, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: The skill implements a standard release engineering workflow for project maintainers. It facilitates release branch creation, version bumping, and documentation synchronization within the NVIDIA NeMo Fabric repository.
  • [COMMAND_EXECUTION]: The skill uses standard development utilities including git, cargo, just, and rg (ripgrep). These tools are used for their intended purposes: version control, build management, and text searching within the local repository context.
  • [INDIRECT_PROMPT_INJECTION]: The skill exhibits an attack surface for indirect prompt injection as it processes untrusted data from repository files.
  • Ingestion points: Reads Cargo.toml, README.md, and various documentation/example files across the docs/, examples/, and adapters/ directories.
  • Boundary markers: None identified; instructions are interpolated directly into the workflow context.
  • Capability inventory: Significant capabilities including git push to the upstream repository, file system modifications via just set-version, and build execution via cargo check and just build-all.
  • Sanitization: No explicit sanitization or filtering of the content read from files is performed before processing.
  • Assessment: While the attack surface exists, the risk is consistent with standard developer tooling for repository maintenance, and the intended use case involves trusted maintainers operating on project source code.
Audit Metadata
Risk Level
SAFE
Analyzed
Oct 3, 2026, 05:20 PM