expert-advice
Pass
Audited by Gen Agent Trust Hub on Sep 14, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill uses the
$ARGUMENTSvariable to incorporate user-supplied input directly into the instructions under the header 'User's assist request:'. - Ingestion points: The
$ARGUMENTSplaceholder inSKILL.mdaccepts arbitrary text from the user. - Boundary markers: Absent. The input is not wrapped in delimiters or accompanied by instructions to treat it as data only.
- Capability inventory: The instructions grant the agent the ability to 'Explore the current codebase' and 'Fetch the Wasp docs', which could be abused if the agent follows commands hidden in the user input.
- Sanitization: No evidence of input validation or sanitization before the content is processed by the LLM.
Audit Metadata