The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill requires the installation of the @t2000/cli package from the NPM registry. This package is an external dependency from a source not identified as a well-known service in the context of this analysis.\n- [REMOTE_CODE_EXECUTION]: The instruction to run npx @t2000/cli init directly downloads and executes code from a remote source at runtime, which is a significant execution vector.\n- [COMMAND_EXECUTION]: The skill's primary functionality relies on executing shell commands, specifically t2000 save <amount>, which runs on the host system.\n- [PROMPT_INJECTION]: The skill is vulnerable to indirect prompt injection because it interpolates the <amount> variable directly into a shell command without sanitization or boundary markers.\n
Ingestion points: The <amount> parameter in the t2000 save command is sourced from user instructions.\n
Boundary markers: Absent. The prompt does not provide delimiters to isolate the input value from the executable command.\n
Capability inventory: The skill uses the t2000 CLI tool, which has the capability to interact with the network and file system to process transactions on the Sui network.\n
Sanitization: Absent. There is no logic to validate that the <amount> input is a numeric scalar, allowing for the possibility of shell metacharacter injection.

t2000-save