skills/jeremylongshore/claude-code-plugins-plus-skills/monitoring-cross-chain-bridges/Gen Agent Trust Hub
monitoring-cross-chain-bridges
Pass
Audited by Gen Agent Trust Hub on Mar 11, 2026
Risk Level: SAFE
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill fetches real-time bridge data from well-known and reputable external services, including DefiLlama (bridges.llama.fi) and official bridge protocol endpoints (wormholescan.io, layerzero.network, across.to). These connections are essential to the skill's primary purpose and target trusted domains.\n- [EXTERNAL_DOWNLOADS]: The transaction tracking and protocol adapter components connect to public blockchain RPC endpoints (such as LlamaRPC, Binance, and Polygon) to verify transaction confirmations on-chain.\n- [PROMPT_INJECTION]: The skill possesses an indirect prompt injection surface typical of monitoring tools that ingest external data.\n
- Ingestion points: User-provided transaction hashes and structured JSON data returned from external bridge APIs.\n
- Boundary markers: There are no explicit delimiters or specific instructions for the agent to ignore embedded instructions in the processed data.\n
- Capability inventory: The skill is granted access to standard file tools (Read, Write, Edit, Grep, Glob) and a restricted Bash tool limited to executing the bridge monitor script.\n
- Sanitization: Input transaction hashes and API responses are processed via standard Python data structures and formatting logic, which provides inherent protection against common injection patterns.\n- [CREDENTIALS_UNSAFE]: The implementation guide references loading API credentials from a local environment file (config/crypto-apis.env). This is a standard security practice for managing user-provided credentials and does not include any hardcoded secrets within the skill's codebase.
Audit Metadata