information-security-manager-iso27001
Pass
Audited by Gen Agent Trust Hub on Sep 4, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection because it ingests untrusted data from external files and interpolates it into the agent's context without adequate boundaries or sanitization.
- Ingestion points: The
scripts/risk_assessment.pyscript ingests data from a user-specified CSV file via the--assetsparameter (lines 122-142), andscripts/compliance_checker.pyingests data via the--controls-fileparameter (lines 135-148). - Boundary markers: The scripts do not use explicit boundary markers or instructions to ignore embedded commands when reading from these CSV files. The content is directly mapped into internal dictionaries and then formatted into Markdown tables or JSON structures.
- Capability inventory: The skill allows for local file writing (scripts save reports to user-defined paths) and instructs the agent to interpret these reports to provide security guidance and gap analysis. A compromised asset name or control status in a CSV file could influence the agent's reasoning or output when it processes the generated report.
- Sanitization: There is no evidence of sanitization, escaping, or validation of the text content within the CSV rows before they are included in the final report output.
Audit Metadata