ln-720-structure-migrator
Pass
Audited by Gen Agent Trust Hub on Mar 23, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill includes a fallback mechanism to fetch reference files from the author's GitHub repository (levnikolaevich/claude-code-skills) via WebFetch. This is a vendor-owned resource used for retrieving structure templates.
- [COMMAND_EXECUTION]: The skill instructs the agent to run standard build commands such as 'npm run build' and 'dotnet build' to verify the success of the restructuring process.
- [PROMPT_INJECTION]: The skill analyzes existing project files and structures, which introduces a surface for indirect prompt injection from untrusted code. It mitigates this by using a structured orchestration workflow and delegating specialized tasks to child agents. Ingestion points: Project structure scan and framework detection in SKILL.md. Boundary markers: None explicitly defined for scanned data. Capability inventory: File system analysis, build command execution, and tool delegation to child agents. Sanitization: Relies on structured delegation logic.
Audit Metadata