grove-issues

Pass

Audited by Gen Agent Trust Hub on Oct 9, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted user-provided text (brain dumps) and uses it to populate titles and bodies of GitHub issues.
  • Ingestion points: User messages containing batches of tasks or ideas are parsed and used in the issue creation process defined in SKILL.md.
  • Boundary markers: The skill uses a structured markdown template and a quoted shell heredoc (cat <<'EOF') to encapsulate the issue body, providing a degree of separation between command instructions and user data.
  • Capability inventory: The skill utilizes gh issue list to search for duplicates and gh issue create to interact with remote repositories.
  • Sanitization: The instructions do not explicitly require the agent to escape or sanitize user-provided text before interpolating it into the shell commands, relying on the agent's formatting and the platform's command execution safety.
  • [COMMAND_EXECUTION]: The skill requires the use of the gh (GitHub CLI) utility and jq for issue management operations. It dynamically constructs shell commands that incorporate user-supplied task descriptions, which is a standard pattern for CLI-based agent skills but presents a common attack surface if the underlying platform does not handle argument escaping correctly.
Audit Metadata
Risk Level
SAFE
Analyzed
Oct 9, 2026, 01:31 AM