copywriting
Pass
Audited by Gen Agent Trust Hub on Sep 21, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and analyze untrusted external data, which creates a vulnerability surface for indirect prompt injection attacks.
- Ingestion points: Data enters the agent context through web search results via Bing (for competitor analysis) and by reading external files containing customer reviews, survey data, and raw interview transcripts.
- Boundary markers: The instructions in SKILL.md and reference files do not specify the use of delimiters (e.g., XML tags or triple quotes) or provide 'ignore embedded instructions' warnings when processing this external research content.
- Capability inventory: The skill utilizes tool invocation for browsing (Bing) and file-read operations (prompt library and research data). It does not exhibit high-privilege capabilities such as arbitrary code execution or network exfiltration.
- Sanitization: The skill lacks explicit sanitization, validation, or filtering logic to identify and neutralize malicious instructions that might be embedded within analyzed customer reviews or web-retrieved content.
Audit Metadata