fancy-build
Pass
Audited by Gen Agent Trust Hub on Oct 9, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTIONDYNAMIC_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted content from various project files (tasks.md, design.md, decisions.md, etc.) which influences agent behavior and triggers command execution. Evidence: Ingestion points include tasks.md, design.md, and visual.md (SKILL.md). No boundary markers or sanitization logic are defined. The capability inventory includes file editing and command execution.
- [COMMAND_EXECUTION]: The skill instructions direct the agent to execute shell commands for verification purposes. Evidence: The agent is instructed to run git log and execute any "Verify command" found in the tasks.md file (SKILL.md).
- [DYNAMIC_EXECUTION]: The agent executes commands that are dynamically retrieved from the tasks.md file during the implementation loop. Evidence: The agent takes a row from tasks.md and runs its associated Verify command (SKILL.md), which is a form of dynamic execution based on local project data.
Audit Metadata