signup-flow-cro
Pass
Audited by Gen Agent Trust Hub on Feb 12, 2026
Risk Level: LOWNO_CODE
Full Analysis
The provided SKILL.md file is a descriptive markdown document. It outlines principles, best practices, and questions for an AI agent to use when advising on signup flow Conversion Rate Optimization (CRO). No executable code, scripts, or external commands are present.
Threat Category Analysis:
- Prompt Injection: No patterns indicative of prompt injection (e.g., 'IMPORTANT: Ignore', 'CRITICAL: Override', role-play injection) were found. The instructions are clear directives for the AI's intended function.
- Data Exfiltration: The skill instructs the AI to read
.claude/product-marketing-context.md. This is a file system access, but it targets a non-sensitive, agent-specific context file, not user credentials or system files. No network operations (e.g.,curl,wget) are present. - Obfuscation: No Base64, zero-width characters, Unicode homoglyphs, or other obfuscation techniques were detected.
- Unverifiable Dependencies: The skill does not attempt to install packages (
npm,pip) or download external scripts. - Privilege Escalation: No commands like
sudo,chmod, or attempts to modify system files or services were found. - Persistence Mechanisms: No attempts to modify shell configurations (
.bashrc), create cron jobs, or establish other persistence mechanisms were detected. - Metadata Poisoning: The metadata block (
name,version,description) is clean and contains no malicious instructions. - Indirect Prompt Injection: While any AI skill that processes user input carries a general risk of indirect prompt injection, this skill itself does not introduce new vectors for this. The instruction to read
.claude/product-marketing-context.mdis for internal context, not arbitrary user input. - Time-Delayed / Conditional Attacks: No conditional logic based on dates, times, usage counts, or environment variables that could trigger malicious behavior was found.
Conclusion: The skill is safe as it is purely informational and does not perform any actions that could compromise the system or exfiltrate data.
Audit Metadata