skills/nulfrog/skills/git-renormalize/Gen Agent Trust Hub

git-renormalize

Pass

Audited by Gen Agent Trust Hub on Jul 24, 2026

Risk Level: SAFECOMMAND_EXECUTIONDATA_EXFILTRATIONPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill provides complex shell pipelines for Bash and PowerShell to identify and checkout specific files. These commands execute local Git operations based on the state of the working directory.
  • [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted input from the local repository (filenames) and pipes them into shell commands. This creates a vulnerability surface where a malicious repository could use specially crafted filenames to execute arbitrary commands when the user runs the provided normalization scripts.
  • Ingestion points: File names are retrieved via git ls-files -m and git diff --name-only in both Bash and PowerShell environments.
  • Boundary markers: None present. Filenames are processed as raw strings.
  • Capability inventory: The skill uses git add --renormalize, git checkout, and git config. These have the capability to modify the local filesystem and Git configuration.
  • Sanitization: Missing. The scripts do not use null-terminators (e.g., git ls-files -z and xargs -0) to safely handle filenames containing spaces, quotes, or shell metacharacters.
  • [DATA_EXPOSURE]: The skill reads the local Git configuration (core.autocrlf) and the contents of the .gitattributes file to diagnose line-ending issues.
Audit Metadata
Risk Level
SAFE
Analyzed
Jul 24, 2026, 02:06 AM