bmad-prfaq
Pass
Audited by Gen Agent Trust Hub on Nov 6, 2026
Risk Level: SAFECOMMAND_EXECUTIONDYNAMIC_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill instructs the agent to execute shell commands using
uv runto invoke local Python scripts (resolve_customization.py,resolve_config.py). These commands are used for configuration loading and workflow step resolution. - [DYNAMIC_EXECUTION]: Utility scripts are executed from computed paths within the user's project structure (
{project-root}/_bmad/scripts/). This pattern assumes that the scripts located within the project's internal directory are safe to execute. - [INDIRECT_PROMPT_INJECTION]: The
artifact-analyzer.mdsub-agent is designed to scan and ingest arbitrary documents from the project workspace to inform the product coaching process. This creates a vulnerability surface where instructions hidden in project files could attempt to manipulate the agent's behavior. - Ingestion points: Project documents, brainstorming reports, and research findings scanned by the
Artifact Analyzeracross{output_folder}and{project-root}. - Boundary markers: The skill relies on sub-agent synthesis (JSON extraction) to separate data from instructions but does not provide explicit 'ignore' delimiters to the main agent when processing findings.
- Capability inventory: The skill possesses significant capabilities, including local shell command execution (
uv run) and file read/write operations across the workspace. - Sanitization: No explicit sanitization or filtering of text content extracted from project documents was observed.
Audit Metadata