mastra
Pass
Audited by Gen Agent Trust Hub on Mar 23, 2026
Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill instructs the agent to run a shell command (ls) to inspect specific local skill directories (e.g., ~/.claude/skills/) for companion skill discovery. While intended for UX improvement, this involves local filesystem inspection.
- [PROMPT_INJECTION]: The skill describes building agents and workflows that process untrusted external data (e.g., user-provided text for summarization or tool inputs). This creates an attack surface for indirect prompt injection.
- Ingestion points: Workflow steps and tool inputs in SKILL.md.
- Boundary markers: Absent in examples.
- Capability inventory: Model generation (agent.generate), workflow execution, and arbitrary tool execution.
- Sanitization: Absent in examples.
Audit Metadata