maintainx-migration-deep-dive
Pass
Audited by Gen Agent Trust Hub on Sep 9, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONDYNAMIC_CONTEXT_INJECTION
Full Analysis
- [DYNAMIC_CONTEXT_INJECTION]: The skill uses the dynamic context execution pattern
!node --version 2>/dev/null || echo 'N/A'SKILL.mdto verify the execution environment at load time. This is a benign use of the feature for compatibility checking. - [INDIRECT_PROMPT_INJECTION]: The skill ingests and processes untrusted data from external sources such as CSV and Excel files in
SKILL.mdandreferences/implementation-guide.md. While the data is processed through structured parsing and mapping logic, the instructions lack explicit boundary markers or warnings to the agent to ignore potential instructions embedded within the migration records. - Ingestion points: CSV parsing logic in
SKILL.md(assess-source.ts,migrate.ts) and Excel parsing examples in the guide. - Boundary markers: None identified in instructions or code samples to delimit untrusted data from system instructions.
- Capability inventory: The skill possesses capabilities for file system writes (reports and logs) and network operations (MaintainX API interaction via curl and Node.js clients).
- Sanitization: The implementation uses field mapping and type transformation, but lacks content-based sanitization for natural language fields that might contain injected instructions.
Audit Metadata