kelly-invest-stock
Pass
Audited by Gen Agent Trust Hub on Aug 25, 2026
Risk Level: SAFEPROMPT_INJECTIONEXTERNAL_DOWNLOADSCOMMAND_EXECUTION
Full Analysis
- [PROMPT_INJECTION]: The skill processes untrusted data from external market sources and user-provided strategy descriptions, creating a surface for indirect prompt injection.
- Ingestion points: External market data retrieved via
stock-sdkand research evidence records stored in thestrategy-reviewsBusabase resource. - Boundary markers: The skill requires manual human maturity labeling (L1, L2, L3) for strategies, and instructions explicitly separate research evidence from automated ranking.
- Capability inventory: Execution is restricted to virtual ledger operations and does not possess access to live brokerage APIs, order placement, or money movement.
- Sanitization: The frontend application (
app.js) implements standard HTML escaping for all user-controllable and external data displayed in the UI. - [EXTERNAL_DOWNLOADS]: The skill downloads
busabase-sdk,stock-sdk, and development tools likeesbuild-wasmfrom the NPM registry. These are well-known packages used for the skill's defined financial research and database synchronization functions. - [COMMAND_EXECUTION]: Includes utility scripts (
setup.mjs,publish_airapp.mjs) used to initialize the Busabase workspace and deploy the AirApp. These scripts execute shell-level logic to interact with external APIs but are not triggered automatically by the agent instructions via dynamic context injection.
Audit Metadata