dependency-track
Pass
Audited by Gen Agent Trust Hub on Sep 17, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted data from a user-configured Dependency-Track server, which creates an attack surface for indirect prompt injection.
- Ingestion points: Untrusted content such as project names, component descriptions, and vulnerability details (titles, descriptions, and recommendations) is retrieved in
scripts/dtrack-findings.sh,scripts/dtrack-projects.sh, andscripts/dtrack-vulnerability.sh. - Boundary markers: The scripts do not implement explicit boundary markers or "ignore instructions" wrappers when presenting this external data to the AI agent.
- Capability inventory: The skill provides significant write-back capabilities, including the ability to audit findings (mark as false positive, exploitable, etc.) via
scripts/dtrack-audit.shand upload SBOM files viascripts/dtrack-bom-upload.sh. - Sanitization: While the scripts use
jqto ensure valid JSON structure, they do not sanitize or filter the natural language content within the retrieved data for potential prompt injection patterns.
Audit Metadata