mass-desire-angle-finder
Pass
Audited by Gen Agent Trust Hub on Apr 14, 2026
Risk Level: SAFEPROMPT_INJECTION
Full Analysis
- [PROMPT_INJECTION]: The skill possesses a surface for indirect prompt injection due to its core functionality of ingesting and analyzing content from untrusted external websites.
- Ingestion points: The agent is instructed to use
web_fetchto retrieve threads from Reddit, Quora, and other forums, as well as reviews from Amazon (SKILL.md). - Boundary markers: The instructions lack delimiters or explicit boundary markers to separate the untrusted data being fetched from the agent's internal reasoning and instructions.
- Capability inventory: The skill utilizes
web_searchandweb_fetchto gather data and generates text-based marketing reports. It does not have access to sensitive system files, hardcoded credentials, or arbitrary code execution capabilities. - Sanitization: There is no mechanism described for sanitizing or validating the retrieved external content before it is processed by the agent.
Audit Metadata