outside-in-testing
Pass
Audited by Gen Agent Trust Hub on Mar 14, 2026
Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill orchestrates the execution of external applications (CLI, TUI, Web, Electron) to verify their behavior. This involves spawning subprocesses and simulating user interactions like keystrokes and clicks.
- [EXTERNAL_DOWNLOADS]: The
scripts/check-freshness.pyscript performs a network request to the GitHub API (api.github.com) to compare the local framework version with the latest release. - [EXTERNAL_DOWNLOADS]: Installation instructions recommend fetching the framework from the vendor's GitHub repository or via the npm registry.
- [PROMPT_INJECTION]: The skill processes structured test scenarios (YAML) that define agent actions, representing an indirect prompt injection surface.
- Ingestion points: Scenario YAML files in the
examples/directory. - Boundary markers: Structured YAML schema for test definitions.
- Capability inventory: Capabilities include process launching, input simulation, and shell command execution across agents.
- Sanitization: The framework employs AI-based 'Comprehension Agents' to validate application output against expected results.
Audit Metadata