transilience-report-style

Pass

Audited by Gen Agent Trust Hub on Jul 30, 2026

Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The smoke test script reference/test_generate_report.py utilizes subprocess.run to invoke the generator script and verify output produced during testing.
  • [EXTERNAL_DOWNLOADS]: The image_flowable function in reference/generate_report.py uses urllib.request to fetch images from URLs provided in the input data for inclusion in the findings cards.
  • [PROMPT_INJECTION]: The skill processes external findings data, representing an indirect injection surface. 1. Ingestion points: report_data.json (specifically the findings and poc arrays). 2. Boundary markers: Absent in data, relying on structured JSON and prose documentation. 3. Capability inventory: Subprocess calls in reference/test_generate_report.py; network read and PDF file-write in reference/generate_report.py. 4. Sanitization: HTML escaping and PII masking via regular expressions in reference/generate_report.py.
Audit Metadata
Risk Level
SAFE
Analyzed
Jul 30, 2026, 04:37 PM
Security Audit — agent-trust-hub — transilience-report-style