saas-architect

Warn

Audited by Gen Agent Trust Hub on Jul 18, 2026

Risk Level: MEDIUMEXTERNAL_DOWNLOADSREMOTE_CODE_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [REMOTE_CODE_EXECUTION]: The skill documentation instructs users or agents to execute npx bmad-method@latest install, which downloads and runs code from the NPM registry. This represents a remote code execution vector from a third-party source not listed as a known trusted provider.\n- [EXTERNAL_DOWNLOADS]: The skill requires the download of the bmad-method package from the NPM registry to support its primary workflow functionality.\n- [PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection through its data ingestion process, which could allow malicious instructions in a product brief to influence the resulting architecture and downstream development tasks.\n
  • Ingestion points: The agent is instructed to ask for and process a 'Product description' and 'Core features' from the user in the 'Understand the brief' step of the architecture design workflow.\n
  • Boundary markers: There are no explicit delimiters or instructions to ignore potential commands embedded within the user-provided brief.\n
  • Capability inventory: The agent possesses the capability to generate comprehensive system designs, API contracts, and 'handoff briefs' that are intended to be processed automatically by subsequent coding agents.\n
  • Sanitization: The skill lacks any instructions for validating, escaping, or sanitizing the content of the ingested product brief before it is incorporated into the architecture documentation.
Audit Metadata
Risk Level
MEDIUM
Analyzed
Jul 18, 2026, 04:19 PM
Security Audit — agent-trust-hub — saas-architect