migrate-content-ia
Pass
Audited by Gen Agent Trust Hub on Oct 2, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- Command Execution (Local): The skill utilizes local command-line tools, including
git,ripgrep, anddocker, to perform repository analysis, identify changes, and validate build integrity. These operations are conducted locally and align with standard development practices for managing documentation infrastructure. - Repository Content Ingestion: The skill involves scanning files within the
content/,data/, andlayouts/directories to identify links and anchors that require updating. Because this process ingests data directly from the repository into the agent's context without explicit boundary markers or sanitization, it presents a potential surface for indirect prompt injection if the source files were to contain instructions intended to influence agent behavior. - Ingestion points: Content files, data configurations, and layout partials within the repository.
- Boundary markers: Not explicitly defined in the provided instructions for data read from files during the search and sweep phases.
- Capability inventory: The agent is authorized to perform file writes and execute shell commands (
git,docker,rg) as part of the migration and validation phases. - Sanitization: No specific sanitization, escaping, or filtering of the ingested file content is specified before the agent processes or presents the data.
Audit Metadata