ralph-copywriter
Pass
Audited by Gen Agent Trust Hub on Oct 15, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and analyze external data from various sources such as blog posts, social media profiles, competitor content, and market trends to generate content and voice profiles. This creates a surface for indirect prompt injection if the external content contains hidden instructions designed to influence the AI's behavior.
- Ingestion points: The skill analyzes existing blog posts, Twitter/LinkedIn profiles, competitor content, and market questions as part of its 'Discover' and 'Research' phases (documented in
SKILL.md). - Boundary markers: No explicit instructions for delimiters or safety warnings regarding embedded instructions in external data are present in the skill definition.
- Capability inventory: The skill is capable of reading and analyzing external content and then generating/writing new content based on that analysis (documented in
SKILL.md). - Sanitization: No sanitization, escaping, or filtering of the ingested external content is specified before it is processed by the agent.
Audit Metadata