aisa
Pass
Audited by Gen Agent Trust Hub on Oct 22, 2026
Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill uses the
aisaCLI to perform search, quote, and call operations. These are standard tool-calling mechanisms provided by the vendor to interact with their platform. - [EXTERNAL_DOWNLOADS]: The skill references a remote MCP endpoint at
https://tools.aisa.one/mcpand documentation ataisa.one. These are verified vendor-owned resources and do not represent unauthorized external communication. - [INDIRECT_PROMPT_INJECTION]: The skill ingests tool metadata and schemas via search results which are external data. While this presents an attack surface, the instructions specifically mitigate risk by directing the agent to use only the structured results provided by the search and schema tools and prohibiting the invention of tool IDs or arguments.
Audit Metadata