ptbs
Pass
Audited by Gen Agent Trust Hub on Oct 7, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONEXTERNAL_DOWNLOADS
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill instructs the agent to utilize an external Model Context Protocol (MCP) server at
https://sui.mcp.kapa.aito retrieve up-to-date Sui documentation. This creates a surface for indirect prompt injection via tool output poisoning, as the agent may process and follow instructions embedded within the documentation returned by this third-party service. - Ingestion points: Documentation results from the MCP tool at
https://sui.mcp.kapa.aias directed inSKILL.md. - Boundary markers: The skill does not define specific delimiters or instructions to ignore embedded directives in external documentation content results.
- Capability inventory: The skill enables the agent to compose, review, and debug Sui blockchain transactions (PTBs), which involve asset transfers and smart contract interactions.
- Sanitization: No sanitization or validation of the documentation content fetched from the remote server is prescribed.
- [EXTERNAL_DOWNLOADS]: The skill references several external resources for documentation and blockchain interaction, including
docs.sui.io,sdk.mystenlabs.com, andfullnode.mainnet.sui.io. These are established vendor resources belonging to the Sui ecosystem and Mysten Labs.
Audit Metadata