skills/xdlkc/rock-cli-skills/rock-cli/Gen Agent Trust Hub

rock-cli

Pass

Audited by Gen Agent Trust Hub on Mar 19, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADSREMOTE_CODE_EXECUTIONCOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill instructs users to install the tool by downloading and executing a shell script from 'http://xrl.alibaba-inc.com/install.sh'. This is a standard installation method for internal tools from well-known technology organizations.
  • [REMOTE_CODE_EXECUTION]: Provides functionality to execute arbitrary commands within a sandbox via 'rockcli sandbox exec' and 'rockcli sandbox attach'. This is the primary feature of the skill for running scripts and debugging.
  • [COMMAND_EXECUTION]: Uses local CLI commands ('rockcli', 'rc') to perform network-based operations and sandbox lifecycle management.
  • [PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection because it reads external data that could contain adversarial content.
  • Ingestion points: Data is ingested from the sandbox through 'log search', 'log tail', and 'history' commands (SKILL.md, references/sandbox.md).
  • Boundary markers: No specific delimiters or safety warnings are provided to prevent the agent from obeying instructions found in these data sources.
  • Capability inventory: The agent can execute commands in a sandbox ('exec') and perform file operations ('upload', 'download').
  • Sanitization: No sanitization or filtering of log/history data is implemented.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 19, 2026, 12:57 PM
Security Audit — agent-trust-hub — rock-cli