skills/tenequm/skills/erc-8004/Gen Agent Trust Hub

erc-8004

Pass

Audited by Gen Agent Trust Hub on Mar 7, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONDATA_EXFILTRATIONPROMPT_INJECTION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill documentation references and utilizes the 'agent0-sdk' package available via the npm registry, which is the official software development kit for the ERC-8004 protocol.
  • [COMMAND_EXECUTION]: The provided implementation guides include standard shell commands for package installation ('npm install agent0-sdk').
  • [DATA_EXFILTRATION]: The skill facilitates necessary communication with external blockchain RPC endpoints, IPFS pinning services (Pinata), and The Graph subgraphs to enable agent registration, reputation management, and multi-chain discovery. These operations target legitimate infrastructure and well-known services associated with the protocol's operation.
  • [PROMPT_INJECTION]: The agent discovery mechanism ingests metadata (names and descriptions) from decentralized registries. While this constitutes an indirect prompt injection surface, it is an inherent property of discovery protocols and is handled as a standard functional risk of agent-to-agent interaction.
  • Ingestion points: Agent registration files (JSON) fetched from IPFS or HTTPS via SDK methods like 'sdk.searchAgents' and 'sdk.getAgent'.
  • Boundary markers: Not explicitly implemented in the provided documentation snippets.
  • Capability inventory: The skill enables on-chain agent registration, feedback submission, wallet verification, and multi-chain discovery queries.
  • Sanitization: No explicit sanitization or filtering logic for descriptive metadata is described within the skill documentation.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 7, 2026, 12:49 AM