bmad-prd

Pass

Audited by Gen Agent Trust Hub on Nov 6, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTIONDYNAMIC_EXECUTIONEXTERNAL_DOWNLOADS
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill processes potentially untrusted external data during its Discovery and Validation phases, creating an indirect prompt injection surface.
  • Ingestion points: The skill ingests data from initiative files, project tickets, epic definitions, and various user-provided inputs like competitive analysis, research, and transcripts (SKILL.md).
  • Boundary markers: To mitigate injection risks, the instructions mandate using subagents to "source-extract" information rather than ingesting entire documents into the parent context (SKILL.md).
  • Capability inventory: The skill can write files to the workspace, execute local Python scripts via uv run, and invoke system openers to display reports.
  • Sanitization: While the workflow uses subagent-based extraction to filter content, it does not detail specific character escaping or sanitization for all interpolated prompt variables.
  • [COMMAND_EXECUTION]: The skill executes platform-specific shell commands to provide the user with a visual view of the generated PRD validation reports.
  • Evidence: The references/validate.md file contains instructions to execute open, xdg-open, or start "" to launch the validation-report.html file in the user's default browser.
  • [DYNAMIC_EXECUTION]: The skill performs dynamic rendering of HTML and Markdown content based on data extracted from PRD reviews.
  • Evidence: The synthesis pipeline reads outputs from various reviewer subagents and replaces placeholders in the assets/validation-report-template.html skeleton to generate a consolidated report (references/validate.md).
  • [EXTERNAL_DOWNLOADS]: The skill references and suggests the download of external components, primarily from the vendor's own repositories.
  • Evidence: The bmod.toml file points to github:bmad-code-org/BMAD-METHOD/skills as a source, and the SKILL.md instructions suggest installing additional skills via npx skills add bmad-code-org/BMAD-METHOD.
Audit Metadata
Risk Level
SAFE
Analyzed
Nov 6, 2026, 09:44 PM