longbridge-quant
Pass
Audited by Gen Agent Trust Hub on Oct 12, 2026
Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSINDIRECT_PROMPT_INJECTIONDYNAMIC_EXECUTION
Full Analysis
- [COMMAND_EXECUTION]: The skill extensively utilizes the
longbridgeCLI to fetch financial data, execute quantitative scripts, and manage account positions. These commands are integral to the vendor's provided quantitative analysis tools. - [EXTERNAL_DOWNLOADS]: Instructions recommend installing standard data science and quantitative libraries via pip, including
scikit-learn,pandas,numpy,statsmodels,scipy,arch,xgboost, andlightgbm. It also references the Navi CLI for script validation from the official Navi documentation. - [INDIRECT_PROMPT_INJECTION]: The skill processes structured external data including K-line history, trade logs, index constituents, and portfolio positions, which are used to generate analysis and trading signals.
- Ingestion points: Market data is retrieved via
longbridge kline,longbridge trades,longbridge constituent,longbridge calc-index, andlongbridge positions(SKILL.md, references/hedging.md). - Boundary markers: No specific delimiters or warnings for the agent to ignore embedded instructions in the ingested data are present.
- Capability inventory: The skill can execute shell commands via the
longbridgeCLI, perform local Python processing, and trigger server-side script execution through thelongbridge quantcommand. - Sanitization: There is no evidence of sanitization or filtering applied to the retrieved market data before it is processed by the agent.
- [DYNAMIC_EXECUTION]: The skill generates Python code skeletons for strategy optimization, volatility modeling, and machine learning tasks (references/strategy-optimizer.md, references/ml-strategy.md). It also supports running Navi and PineScript code server-side using the
longbridge quant runcommand (references/quant-cli.md).
Audit Metadata