tres-explorer-tx-to-ledger
Pass
Audited by Gen Agent Trust Hub on Aug 1, 2026
Risk Level: SAFE
Full Analysis
- Untrusted Data Handling: The skill explicitly includes security instructions to treat data fetched from blockchain explorers as untrusted text. It specifically warns against following any instructions embedded in the external pages, which effectively mitigates indirect prompt injection risks.
- Domain Validation: It implements a strict allowlist for supported blockchain explorer domains (e.g., etherscan.io, bscscan.com) and refuses to process requests from unknown sources.
- Human-in-the-Loop Verification: Before executing any GraphQL mutations that modify the ledger, the skill requires the agent to present a structured summary of the transaction data and obtain explicit user approval.
- Prompt Injection Defense: While the prompt injection detector flagged instructions related to ignoring instructions, these are actually protective measures. The skill instructs the agent to ignore directives within the untrusted external data to ensure it remains focused on its primary task of data extraction.
Audit Metadata