The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill instructs the user and agent to install the official @onekeyfe/cli package from the NPM registry. This is an expected vendor resource for the OneKey ecosystem.
[COMMAND_EXECUTION]: The skill relies on executing local CLI commands through the onekey tool. Specifically, it uses security-audit, security-simulate, and device-verify to perform its core functions. These commands are restricted to the vendor's provided utility.
[PROMPT_INJECTION]: The skill analyzes untrusted blockchain data such as token contract addresses and transaction calldata, which constitutes an indirect prompt injection surface.
Ingestion points: Untrusted contract addresses, mint addresses, and hex-encoded transaction data are passed to the agent for auditing in SKILL.md.
Boundary markers: The instructions in references/common.md establish "Hard Stops" and chain-mismatch validation to prevent the agent from proceeding with unsafe data.
Capability inventory: The agent uses CLI-based auditing and simulation tools that interface with external blockchain data.
Sanitization: The skill includes explicit instructions to identify and flag malicious features such as "hidden mints," "blacklist controls," and "proxy upgrades" rather than executing the instructions contained within the analyzed data.

onekey-security