testing
Pass
Audited by Gen Agent Trust Hub on Oct 31, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTIONDYNAMIC_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill ingests data from external sources including browser content via
agent-browserand BDD.featurefiles. - Ingestion points:
SKILL.md(Scenario-Driven Business Tests) andreferences/browser.md(Screenshot Evidence). - Boundary markers: Instructions explicitly warn that no model should interpret scenarios or supply bindings during execution and emphasize manual review.
- Capability inventory: Shell commands for test runners (
vitest,pytest,cargo) and browser management (playwright-cli,pgrep,kill). - Sanitization: There are no explicit instructions for sanitizing inputs before they are interpolated into shell command templates.
- [COMMAND_EXECUTION]: The skill provides shell command templates for running tests and managing browser processes.
- Evidence: Templates like
vitest run --reporter=verbose -t "<test name>"andpgrep -fl 'remote-debugging-pipe'are intended for agent execution. - [DYNAMIC_EXECUTION]: The skill describes a workflow involving the deterministic generation of executable tests from business scenarios.
- Evidence:
SKILL.mdoutlines a process for generating and running tests based on.featurefiles.
Audit Metadata