multiversx-fix-verification
Pass
Audited by Gen Agent Trust Hub on Feb 17, 2026
Risk Level: SAFE
Full Analysis
- SAFE (SAFE): No malicious patterns or behaviors were identified. The skill correctly identifies common security pitfalls like partial fixes, reentrancy-induced DoS, and integer overflows, providing legitimate remediation strategies.
- COMMAND_EXECUTION (SAFE): The skill references standard development tools such as 'cargo test', 'cargo clippy', and 'sc-meta' for local testing and verification. These are expected within the context of smart contract development and do not involve untrusted remote sources.
- Indirect Prompt Injection (SAFE): The skill's primary function is to process external data (vulnerability reports and code patches). While this creates an ingestion surface, the skill is instructional in nature and does not perform unsafe operations on this data. 1. Ingestion points: Review of code modifications and vulnerability reports. 2. Boundary markers: Not explicitly required for instructional content. 3. Capability inventory: Local test execution via 'cargo' and 'sc-meta'. 4. Sanitization: Not applicable for manual review guidelines.
Audit Metadata