btst-testing
Pass
Audited by Gen Agent Trust Hub on Oct 9, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill manages unit and end-to-end testing workflows, which involve processing source code, test specifications, and environment data. This creates a potential surface where an agent could be influenced by malicious instructions embedded in the project files it is tasked to test.\n
- Ingestion points: Test files (e.g., getters.test.ts, smoke.chat.spec.ts) and environment configuration files (.env).\n
- Boundary markers: There are no specific instructions or delimiters provided to separate test data from agent instructions.\n
- Capability inventory: The skill utilizes shell execution (pnpm, bash), process management (kill), and file system access for project setup and cleanup.\n
- Sanitization: Standard testing logic is provided, but there is no validation or sanitization of the content within the files being processed.\n- [COMMAND_EXECUTION]: The skill contains various shell commands and scripts necessary for the framework's development lifecycle.\n
- Evidence: Commands such as pnpm test, pnpm codegen:e2e, and local setup scripts like bash scripts/codegen/setup.sh are used for building and testing the application.\n
- Evidence: The use of kill $(lsof -ti:3006) is included to manage local network ports, which is a standard utility command for developers.
Audit Metadata