feature-builder
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 presents an attack surface where untrusted instructions in a feature description could influence the multi-agent generation pipeline.\n
- Ingestion points: User-supplied descriptions in
/feature-buildand project files analyzed during Phase 1.\n - Boundary markers: The instructions do not specify delimiters or warnings to isolate user input from agent prompts.\n
- Capability inventory: Significant file system write access for generating source code and tests, coupled with command execution for build verification.\n
- Sanitization: No sanitization or filtering of input descriptions is described before they are processed by the sub-agents.\n- [COMMAND_EXECUTION]: The skill executes platform-specific build and test commands like
./gradlewandxcodebuild.\n- [DYNAMIC_EXECUTION]: The skill's primary function involves generating source code and tests at runtime, followed by their compilation and execution. This involves the dynamic assembly and loading of executable content.
Audit Metadata