data-flow-mapping
Pass
Audited by Gen Agent Trust Hub on Oct 2, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill analyzes project source code, database schemas, and migration files, which serves as an ingestion point for untrusted content. Maliciously crafted comments or code within these files could potentially attempt to influence the agent's behavior. The ingestion occurs through the
scripts/inventory-data-model.mjshelper and direct AI inspection of codebases, without defined boundary markers or explicit sanitization instructions for the analyzed data. - [COMMAND_EXECUTION]: The skill executes a local utility script,
scripts/inventory-data-model.mjs, using Node.js to perform a static analysis of database models and project dependencies to identify PII-related columns. The execution is triggered by the commandnode scripts/inventory-data-model.mjs --root .as described in the skill instructions.
Audit Metadata