pov-vlog-presence-by-2zdsi5
Pass
Audited by Gen Agent Trust Hub on Oct 9, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSINDIRECT_PROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill instructs the agent to visit external websites to inspect media files and documentation from domains including goodcase.ai and x.com.
- [INDIRECT_PROMPT_INJECTION]: The workflow requires the agent to read and summarize content from external sources and the local
references/cases.mdfile to generate video prompts. This process lacks clear boundary markers or sanitization, creating a surface where malicious instructions embedded in the external data could override the agent's behavior. - Ingestion points: The skill ingests information from the
references/cases.mdfile and external URLs via browser or media analysis tools. - Boundary markers: No specific delimiters or safety instructions are defined to separate external data from the skill's internal logic.
- Capability inventory: The agent can access external websites and generate detailed video prompts based on the retrieved information.
- Sanitization: The skill does not perform any validation or filtering of the data retrieved from external sources before using it in the prompt generation process.
Audit Metadata