threejs-qa-release
Pass
Audited by Gen Agent Trust Hub on Sep 19, 2026
Risk Level: SAFEDYNAMIC_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [DYNAMIC_EXECUTION]: The
scripts/inspect-threejs-canvas.mjsscript dynamically resolves and imports Node.js dependencies (@playwright/testandpngjs) from the local project's environment usingimport()with computed file URLs. While this is a common architectural pattern for dev-tooling to avoid bundling heavy dependencies, it involves runtime code loading from local paths. - [INDIRECT_PROMPT_INJECTION]: The skill represents a surface for indirect prompt injection as it ingests untrusted data from the browser environment during testing.
- Ingestion points: Ingests browser console output, page errors, and game-specific diagnostics via
window.__THREE_GAME_DIAGNOSTICS__inscripts/inspect-threejs-canvas.mjs. - Boundary markers: No explicit boundary markers or instruction-guarding delimiters are used when processing browser-sourced strings.
- Capability inventory: The skill has capabilities for file system writes (writing report artifacts), browser automation (via Playwright), and network navigation.
- Sanitization: There is no evidence of sanitization or validation of the data retrieved from the browser context before it is saved to JSON reports or printed to the console.
Audit Metadata