longbridge-hedging

Pass

Audited by Gen Agent Trust Hub on May 15, 2026

Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill instructs the agent to use the longbridge CLI to perform various tasks such as fetching k-line data (longbridge kline), option chains (longbridge option), and portfolio positions (longbridge positions). This is the intended primary functionality of the skill using vendor-provided tools.\n- [PROMPT_INJECTION]: The skill exhibits an indirect prompt injection surface by interpolating untrusted user inputs (like stock symbols and benchmarks) directly into shell command templates.\n
  • Ingestion points: User-provided variables <SYMBOL> and <BENCHMARK> in SKILL.md.\n
  • Boundary markers: Absent; inputs are not delimited or restricted to prevent command injection.\n
  • Capability inventory: Use of shell-based CLI commands across multiple scripts to access market and sensitive portfolio data.\n
  • Sanitization: There are no instructions for validating or sanitizing user inputs before they are executed in the shell.
Audit Metadata
Risk Level
SAFE
Analyzed
May 15, 2026, 11:56 AM