bmad-architecture

Pass

Audited by Gen Agent Trust Hub on Aug 30, 2026

Risk Level: SAFECOMMAND_EXECUTIONDYNAMIC_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill invokes local Python scripts (such as memlog.py, resolve_config.py, and lint_spine.py) using uv run. These scripts are used to manage the session's memory log and perform mechanical validation of the architecture documents.
  • [DYNAMIC_EXECUTION]: The skill configuration allows for the execution of arbitrary activation and completion steps defined in customize.toml. Additionally, it employs a 'Reviewer Gate' that dispatches parallel subagents to evaluate the generated content.
  • [INDIRECT_PROMPT_INJECTION]: The skill ingests data from external sources like existing codebases and specification files, creating a potential vector for indirect prompt injection.
  • Ingestion points: Reads from SPEC.md, project source code, and user-provided raw ideas.
  • Boundary markers: There are no explicit instructions to use delimiters or ignore embedded prompts when processing these external files.
  • Capability inventory: The skill has the ability to write files and execute local commands via uv run.
  • Sanitization: No specific sanitization or filtering mechanisms are defined for the ingested project data.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 30, 2026, 10:42 PM
Security Audit — agent-trust-hub — bmad-architecture