azure-devops-create-work-item
Pass
Audited by Gen Agent Trust Hub on Nov 11, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted user notes and repository files to draft work items, which presents a surface for indirect prompt injection.
- Ingestion points: User-supplied notes and files within the local repository searched during the repository pass in SKILL.md.
- Boundary markers: The instructions explicitly require the agent to redact confidential information and record ambiguities rather than inventing data.
- Capability inventory: The skill can create directories and write Markdown files locally, and it has the capability to read repository files.
- Sanitization: Instructions in SKILL.md mandate generalizing or redacting sensitive identifiers, URLs, and environment names.
Audit Metadata