gsd-review
Pass
Audited by Gen Agent Trust Hub on Jul 22, 2026
Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill invokes several external command-line tools including Gemini, Claude, Codex, Qwen, and Cursor using the Bash tool to perform its primary function of plan review.
- [DATA_EXFILTRATION]: Project phase plans are transmitted to external AI service providers for analysis. This is a core feature of the skill and involves well-known technology services.
- [PROMPT_INJECTION]: An indirect prompt injection surface is present as the skill processes external phase plans and passes them to other AI models. 1. Ingestion points: Project phase plans are read from local files. 2. Boundary markers: No explicit delimiters or instructions are provided to the external CLIs to ignore embedded instructions within the plans. 3. Capability inventory: The skill can execute shell commands via Bash and write results to the file system using the Write tool. 4. Sanitization: No content sanitization or validation is specified before passing data to the review tools.
Audit Metadata