matrixscan-ar-net-android
Pass
Audited by Gen Agent Trust Hub on Oct 22, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill instructs the agent to fetch information from external websites (NuGet and Scandit documentation) to provide accurate SDK versioning and API usage. This ingestion of external data establishes a potential surface for indirect prompt injection.
- Ingestion points:
references/integration.mdandreferences/migration.mdtrigger WebFetch calls to external documentation and package registry pages. - Boundary markers: The skill does not provide specific instructions to delimit or isolate the fetched content from the rest of the prompt context.
- Capability inventory: The agent has the ability to write to the project's filesystem and access license keys via a connected MCP server.
- Sanitization: There are no instructions provided for sanitizing or filtering the content retrieved from external sources.
- [SAFE]: The skill employs secure practices for credential management, explicitly instructing the agent not to search the environment for secrets and requiring that license keys be handled through secure MCP tools or manual user input with masking. All identified network destinations are well-known services or official vendor domains.
Audit Metadata