bmad-party-mode
Pass
Audited by Gen Agent Trust Hub on Sep 17, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTIONDYNAMIC_EXECUTION
Full Analysis
- [COMMAND_EXECUTION]: The script
scripts/resolve_party.pyutilizessubprocess.runto execute internal framework scripts (resolve_config.pyandresolve_customization.py) for roster and configuration resolution. - [INDIRECT_PROMPT_INJECTION]: The skill provides a workflow in
references/create-party.mdto distill personas from user-provided data sources (e.g., spreadsheets or notes). This data is processed to create system instructions for subagents, forming a surface for indirect prompt injection. - Ingestion points:
references/create-party.md(Distill from source data). - Boundary markers: None identified in the distillation instructions.
- Capability inventory: Command execution for internal utilities and file system access for memory logging.
- Sanitization: No sanitization is specified for the distilled persona text.
- [DYNAMIC_EXECUTION]: The skill dynamically assembles prompts, including personas and scene context, to brief and manage subagents and agent teams at runtime across conversation rounds.
Audit Metadata