paywall-upgrade-cro
Pass
Audited by Gen Agent Trust Hub on Jul 27, 2026
Risk Level: SAFEPROMPT_INJECTION
Full Analysis
- [SAFE]: The skill consists of markdown files and a JSON evaluation file designed to provide strategies for conversion rate optimization. No malicious patterns, obfuscation, credential harvesting, or unauthorized resource access attempts were detected.
- [PROMPT_INJECTION]: The skill contains instructions to ingest external context from
.agents/product-marketing-context.mdor.claude/product-marketing-context.md, which constitutes a surface for indirect prompt injection. However, the risk is negligible as the skill lacks high-risk capabilities to exploit. - Ingestion points:
.agents/product-marketing-context.mdor.claude/product-marketing-context.md(as specified in SKILL.md). - Boundary markers: None identified.
- Capability inventory: None (the skill provides text-based advice and has no network, file-write, or shell execution capabilities).
- Sanitization: None identified.
Audit Metadata