ln-25-architecture-diagram-builder
Pass
Audited by Gen Agent Trust Hub on Oct 26, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill serves as a surface for indirect prompt injection because its primary function requires processing potentially untrusted data from a repository.
- Ingestion points: According to the tool routing and checklist, the skill ingests data from repository files, runtime wiring, infrastructure-as-code (IaC) files, contracts, and other existing artifacts.
- Boundary markers: The skill contains instructions to "Define the evidence boundary and label user-supplied or proposed elements separately" and to label unverified elements, which helps distinguish between different levels of trust.
- Capability inventory: The skill is authorized to use language intelligence and dependency tools to inspect repositories and has the capability to write Mermaid or ASCII diagram files to specific documentation paths (e.g.,
docs/architecture/diagrams/). - Sanitization: The skill mandates summarizing results while keeping the full output of tools outside the agent's context, providing a layer of protection against processing large amounts of potentially malicious raw data directly.
Audit Metadata