triage
Pass
Audited by Gen Agent Trust Hub on Sep 15, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill analyzes untrusted data from git diffs and source code files to generate its report.
- Ingestion points: The skill reads file contents using the
Readtool and consumes output fromgit diffcommands in theSKILL.mdfile. - Boundary markers: The instructions do not specify the use of delimiters or explicit 'ignore embedded instructions' markers to wrap the processed code content.
- Capability inventory: Capabilities are limited to
git(diff, log, status, merge-base, rev-parse),Read,Grep, andGlob. It lacks network access and file-writing capabilities, significantly limiting the potential impact of an injection. - Sanitization: No specific content sanitization or validation is performed on the ingested data before analysis.
- [COMMAND_EXECUTION]: The skill performs shell-based git operations using branch names provided as arguments. The execution environment is restricted to a specific allowlist of git subcommands via the YAML frontmatter configuration.
Audit Metadata