bmad-create-product-brief
Pass
Audited by Gen Agent Trust Hub on Mar 14, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: No malicious patterns, obfuscation, or unauthorized network activity detected. The skill uses a structured 'step-file architecture' to maintain disciplined execution and context management.
- [PROMPT_INJECTION]: No direct prompt injection, safety filter bypass, or 'jailbreak' attempts were identified. The instructions use strict mandatory execution rules to guide the agent as a facilitator.
- [DATA_EXFILTRATION]: The skill accesses local project directories such as
{planning_artifacts}and{product_knowledge}to discover relevant project documentation. There are no patterns suggesting data exfiltration or access to sensitive system paths (e.g., SSH keys, credentials). - [COMMAND_EXECUTION]: No dangerous command execution or shell spawning was detected. The skill relies on standard file operations (read/append) within the agent's environment.
- [PROMPT_INJECTION]: Indirect Prompt Injection Surface: The skill ingests external content from project-related files (e.g.,
*research*.md,*brainstorming*.md) instep-01-init.md. While this is a necessary function for creating a product brief, it represents a surface where malicious instructions in those files could influence agent behavior. Boundary markers are absent in the prompt templates, and there is no explicit sanitization of the ingested data. However, the agent's capabilities are limited to document generation and calling other verified internal skills.
Audit Metadata