test-interactions
Pass
Audited by Gen Agent Trust Hub on Oct 2, 2026
Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSINDIRECT_PROMPT_INJECTION
Full Analysis
- Command Execution: The skill coordinates several shell-based operations using
npx iwsdkandnpmto initialize a development server and execute a series of interaction tests. These are standard procedures for integration testing within a local development workspace. - External Resource Fetching: Dependency management is handled via
npm, which retrieves required libraries from the public npm registry. This is an expected step in preparing the environment for the test suites. - Temporary File Management: Log files and visual confirmation screenshots are directed to the
/tmpdirectory. This approach is a routine way to handle non-persistent diagnostic data generated during a test run. - Indirect Prompt Injection Surface: The skill involves the agent parsing and verifying JSON output from tool calls (Ingestion points:
npx iwsdkoutput in SKILL.md). While the instructions do not specify explicit boundary markers, the capabilities are restricted to the local test environment (Capability inventory:iwsdkCLI,npm,/tmpaccess). The structured nature of JSON provides an inherent validation layer (Sanitization: Inherent structure).
Audit Metadata