argent-qa-e2e
Pass
Audited by Gen Agent Trust Hub on Oct 23, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to process external, untrusted content such as 'test cases, tickets, or acceptance criteria' to generate executable E2E tests.
- Ingestion points: Untrusted data enters the agent context via user-provided documentation, tickets, or requirements used to define the test contract.
- Boundary markers: The instructions do not specify any delimiters or safety warnings to ignore instructions embedded within the input data.
- Capability inventory: The skill utilizes
script:for local command execution,e2e-executefor test execution, ande2e-add-scriptfor creating setup/cleanup routines. This combination allows for arbitrary command execution on the host machine. - Sanitization: There is no evidence of sanitization or validation of the input data before it is translated into executable actions or scripts.
- [COMMAND_EXECUTION]: The skill explicitly supports and instructs the agent on how to use the
script:directive to run local setup or cleanup commands on the machine. While this is part of the intended QA automation workflow, it provides a direct mechanism for code execution if the agent is influenced by malicious input.
Audit Metadata