end-to-end-testing
Pass
Audited by Gen Agent Trust Hub on Sep 15, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill provides automated capabilities to parse and aggregate external test outputs and logs, which creates a data processing vulnerability surface.
- Ingestion points: External logs and markdown test files within the target input directories are processed by
scripts/generate-html-report.pyandscripts/validate-test-data.sh. - Boundary markers: The prompt instructions lack strict boundary markers or instructions telling the agent to ignore executable instructions inside raw log data.
- Capability inventory: Local script execution including Python and Bash operations to read, move, and generate local files.
- Sanitization: The
generate-html-report.pyscript performs raw text substitution of log data directly into the HTML viewer template without HTML escaping, which could allow malicious script code embedded inside an application log to render arbitrarily in the viewer.
Audit Metadata