sui-sdk-js
Pass
Audited by Gen Agent Trust Hub on Sep 16, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONEXTERNAL_DOWNLOADS
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill facilitates the ingestion of untrusted data from the Sui blockchain, such as Move object content and transaction metadata. This creates a potential surface for indirect prompt injection if an attacker crafts malicious on-chain data to influence the agent's logic. However, the skill encourages structured parsing using BCS and the Core API, which mitigates simple schema confusion.
- Ingestion points: Ingestion of object content and transaction data via client.core methods in SKILL.md and references/queries.md.
- Boundary markers: No specific boundary markers are recommended for untrusted data processed by the LLM.
- Capability inventory: Capability to sign and execute transactions and Move calls is documented in SKILL.md.
- Sanitization: Validation helpers are mentioned in references/advanced.md but specific LLM-focused sanitization is absent.
- [EXTERNAL_DOWNLOADS]: The skill references official Mysten Labs packages including @mysten/sui and associated signer libraries. These are standard dependencies for Sui blockchain development and follow well-known industry patterns.
Audit Metadata