config-schema-setup
Fail
Audited by Gen Agent Trust Hub on Feb 15, 2026
Risk Level: HIGHEXTERNAL_DOWNLOADSREMOTE_CODE_EXECUTION
Full Analysis
- EXTERNAL_DOWNLOADS (HIGH): The skill directs the agent to use
curlto fetch content fromhttps://fullstackrecipes.com/api/recipes/config-schema-setup. This domain is not a recognized trusted source. - REMOTE_CODE_EXECUTION (HIGH): Although the
curlcommand does not directly pipe to a shell, the agent is explicitly told to 'fetch the recipe' for configuration setup. In the context of AI agents, recipes typically contain code or instructional steps that the agent then applies to the local environment, effectively executing remote content. - INDIRECT PROMPT INJECTION (HIGH): The skill's primary function relies on ingesting untrusted data from an external URL. If an attacker modifies the content at that URL, they can inject malicious instructions that the agent will follow, potentially leading to data exfiltration or unauthorized system modifications.
Recommendations
- AI detected serious security threats
Audit Metadata