browser-workflow-to-playwright
Pass
Audited by Gen Agent Trust Hub on Mar 15, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill performs local file transformations by reading markdown workflows and writing Playwright test scripts. No external network operations, credential exposures, or suspicious command executions were detected.
- [INDIRECT_PROMPT_INJECTION]: The skill processes content from
/workflows/browser-workflows.md, which is a user-controlled source. - Ingestion points: The skill reads workflow names and steps from
/workflows/browser-workflows.mdto drive codebase searches and code generation. - Boundary markers: No explicit boundary markers or 'ignore' instructions are used when passing workflow content to the subagents.
- Capability inventory: The skill can perform codebase searches via Grep and write to the local file system (e2e/browser-workflows.spec.ts).
- Sanitization: The skill does not explicitly sanitize the markdown content before processing. While malicious instructions in the markdown could theoretically influence the generated code, the risk is limited to functional errors in the output file, which requires user review.
Audit Metadata