skills/aomi-labs/skills/aomi-transact/Gen Agent Trust Hub

aomi-transact

Pass

Audited by Gen Agent Trust Hub on Mar 13, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONPROMPT_INJECTIONCREDENTIALS_UNSAFE
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill verifies and installs the @aomi-labs/client and viem packages from the npm registry if they are not detected in the environment.\n- [COMMAND_EXECUTION]: Uses Bash to execute Aomi CLI tools for session management, transaction review, and signing operations.\n- [PROMPT_INJECTION]: The skill processes user-supplied messages to construct blockchain transactions, creating an indirect prompt injection surface.\n
  • Ingestion points: The <message> parameter in the aomi chat command (SKILL.md).\n
  • Boundary markers: Absent; the skill does not use delimiters or instructions to ignore embedded commands within the user message.\n
  • Capability inventory: The aomi sign command (SKILL.md) allows for the signing and broadcasting of arbitrary transactions to the network.\n
  • Sanitization: No input sanitization or validation logic is specified for the chat messages.\n- [CREDENTIALS_UNSAFE]: The instructions direct users to provide private keys as command-line arguments (--private-key <hex-key>), which is a security risk as keys may be logged in shell history files (e.g., ~/.bash_history).
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 13, 2026, 12:01 AM