handoff-to-claude-code
Pass
Audited by Gen Agent Trust Hub on Oct 7, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill ingests and analyzes HTML, CSS, and JS prototype files from the project environment to generate UI documentation. This creates a surface where untrusted data could theoretically influence the AI's output. Ingestion point: project design files. Boundary markers: The AI is instructed to label files as design references, but no specific sanitization instructions are provided. Capabilities include filesystem writes and use of the present_fs_item_for_download tool.
- [COMMAND_EXECUTION]: The skill uses directory creation and file management commands to assemble the handoff bundle. These operations are localized to the workspace and are essential to the skill's core functionality of packaging assets for the user.
Audit Metadata