signup-flow-cro
Pass
Audited by Gen Agent Trust Hub on Feb 26, 2026
Risk Level: SAFE
Full Analysis
- [PROMPT_INJECTION]: The skill instructs the agent to adopt a persona as an expert in signup flow optimization. These instructions are focused on task performance and do not contain directives to bypass safety filters, reveal system prompts, or disregard core instructions.
- [DATA_EXFILTRATION]: There are no network operations, API calls, or patterns indicating data exfiltration. The skill mentions reading a project-specific context file (
.claude/product-marketing-context.md), which is a standard method for providing local context to an agent and does not involve accessing system-sensitive files. - [COMMAND_EXECUTION]: The skill consists exclusively of markdown documentation and instructions. It does not contain any shell scripts, command-line instructions, or mechanisms to execute arbitrary code.
- [EXTERNAL_DOWNLOADS]: No external URLs, remote scripts, or package manager commands were detected in the file.
- [DATA_EXPOSURE]: The skill does not contain hardcoded credentials, API keys, or links to sensitive internal infrastructure.
Audit Metadata