longbridge-watchlist

Pass

Audited by Gen Agent Trust Hub on Oct 12, 2026

Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTIONEXTERNAL_DOWNLOADS
Full Analysis
  • [COMMAND_EXECUTION]: The skill uses the longbridge CLI to perform all operations, including reading and modifying user watchlists, price alerts, and community lists. The skill includes a mandatory two-step confirmation protocol (Preview then Wait for explicit confirmation) before executing any mutating commands to prevent accidental changes.
  • [EXTERNAL_DOWNLOADS]: Instructions direct users to install the longbridge-terminal tool from the vendor's GitHub repository (github.com/longportapp/longbridge-terminal) if the command is not found.
  • [INDIRECT_PROMPT_INJECTION]: The skill processes data from external sources, specifically community stock lists (Sharelists), which could theoretically contain malicious instructions.
  • Ingestion points: CLI output from longbridge sharelist popular and longbridge sharelist detail (referenced in references/sharelist.md).
  • Boundary markers: The skill uses explicit protocols for handling data but lacks strict delimiters for external content within the LLM instructions.
  • Capability inventory: Execution of longbridge CLI commands for list management, alerts, and watchlists.
  • Sanitization: The skill requests structured JSON output (--format json) for commands, which helps in safely parsing the data.
Audit Metadata
Risk Level
SAFE
Analyzed
Oct 12, 2026, 12:25 PM