create-ai-tool
Pass
Audited by Gen Agent Trust Hub on Feb 20, 2026
Risk Level: SAFEPROMPT_INJECTIONNO_CODE
Full Analysis
- [PROMPT_INJECTION] (LOW): The skill identifies a workflow for ingesting user requirements to generate other AI components, creating a surface for indirect prompt injection. A user could provide malicious instructions meant to be embedded in the generated tool.
- Ingestion points: User input provided during the tool selection process (SKILL.md).
- Boundary markers: None present; the instructions do not explicitly warn the AI to ignore instructions within the user requirements.
- Capability inventory: The skill uses routing to call
/create-skilland/create-agentcommands. - Sanitization: No sanitization logic is present in the markdown guidance.
- [NO_CODE] (SAFE): The skill consists entirely of markdown files and metadata. There are no associated scripts (.py, .js, .sh), which eliminates the risk of direct command execution or malware delivery.
Audit Metadata