maestro-new-track
Pass
Audited by Gen Agent Trust Hub on Mar 3, 2026
Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill orchestrates local project setup by executing standard system commands, including
mkdirfor directory creation andgitfor version control management. It also conditionally invokesbr(Beads), a project state initialization tool, when detected in the runtime environment. These commands are used for their intended administrative purposes and rely on internal or sanitized identifiers. - [PROMPT_INJECTION]: The skill features an indirect prompt injection surface as it interpolates user-provided requirements and interview responses into implementation plans that could guide future autonomous agent actions. This is effectively mitigated through human-in-the-loop security gates (Steps 8 and 9), which require explicit user review and approval of the generated specification and plan before they are finalized or committed. Additionally, the skill applies structural formatting (snake_case and date-stamping) to user-derived track IDs to prevent path injection vulnerabilities.
Audit Metadata