agent-browser-issue
Pass
Audited by Gen Agent Trust Hub on Oct 6, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill instructs the agent to collect data from potentially untrusted web pages to populate GitHub issue reports, creating a surface for indirect prompt injection.
- Ingestion points: Browser page content, URLs, and error blocks collected from external websites during automation failures.
- Boundary markers: The instructions provide a structured Markdown template for the issue body, but do not include explicit delimiters or instructions to ignore malicious directives that might be embedded in the collected web content.
- Capability inventory: The skill requires the capability to write to external services (creating GitHub issues) based on the processed untrusted data.
- Sanitization: There is no mention of sanitization, filtering, or escaping of the collected content before it is interpolated into the GitHub issue template.
Audit Metadata