auditing
Pass
Audited by Gen Agent Trust Hub on Sep 30, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted data from target WordPress sites during performance and accessibility audits. An attacker-controlled site could embed instructions within its HTML, CSS, or metadata to influence the agent's report or behavior.
- Ingestion points: Audited site content, source code, and rendered output retrieved through tools like
take_screenshotandinspect_design. - Boundary markers: Absent; no specific delimiters or "ignore" instructions are defined for processing site data.
- Capability inventory: The skill utilizes
take_screenshot,need_for_speed(performance diagnostics),inspect_design, andrecord_workflow_event. - Sanitization: No explicit sanitization or filtering of the audited site's content is implemented before interpretation.
Audit Metadata