create-issue-tree
Pass
Audited by Gen Agent Trust Hub on Oct 6, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [SAFE]: The skill performs strict input validation using regular expressions for all user-provided arguments, such as --granularity and --root numbers, ensuring they conform to expected formats and preventing command injection.
- [SAFE]: Shell operations are executed securely, utilizing variable quoting and temporary files with automated cleanup via trap to avoid data leaks.
- [SAFE]: The logic includes a 'fail-closed' check that verifies the absence of template placeholders in the final generated issue body before performing any write operations to GitHub.
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted data from GitHub issue titles, creating an indirect prompt injection surface. Ingestion points include
scripts/merge-root-body.shandscripts/create-root-body.shwhich fetch issue data usinggh apiandgh issue viewcommands. These external titles are formatted into Markdown table rows (boundary markers) within the tracking issue. The skill possesses the capability to create and edit issues via the gh CLI. Sanitization is performed by the CELL helper inscripts/tree-lib.sh, which escapes Markdown delimiters (|) and newlines to protect the structural integrity of the generated tables.
Audit Metadata