longbridge-catalyst-radar
Pass
Audited by Gen Agent Trust Hub on Apr 29, 2026
Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill extensively uses the
longbridgeCLI tool to retrieve financial data. It constructs shell commands using user-provided inputs like stock symbols and market codes (e.g.,longbridge quote {symbol}.{market}). - [DATA_EXFILTRATION]: The skill accesses sensitive user information, including watchlist details and portfolio positions, via the
longbridge watchlistandStock Positions API. This behavior is central to the skill's primary purpose of providing investment briefings and relies on the vendor's own infrastructure. - [PROMPT_INJECTION]: The skill is vulnerable to indirect prompt injection. It ingests untrusted data from external sources such as news headlines and regulatory filings through the
longbridge newsandlongbridge filing detailcommands. The instructions lack explicit boundary markers or sanitization procedures to prevent the agent from potentially executing instructions embedded within these external financial documents. Evidence found inreferences/longbridge-api-map.md(ingestion) andSKILL.md(execution capabilities).
Audit Metadata