vision-alignment-foresight

Pass

Audited by Gen Agent Trust Hub on Jul 30, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADSPROMPT_INJECTION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill references and fetches guidelines from trusted and well-known sources, including official documentation for the OpenAI Agents SDK, the Model Context Protocol (Anthropic), and technical specifications from Mintlify and ArXiv.
  • [PROMPT_INJECTION]: The skill possesses a surface for indirect prompt injection as it is designed to analyze untrusted content from the local repository environment.
  • Ingestion points: The skill workflow requires reading North Star documents, Architecture Decision Records (ADRs), configuration files, and source code within the target repository (SKILL.md, workflow step 2).
  • Boundary markers: There are no instructions defining specific delimiters or safety boundaries for the content being analyzed.
  • Capability inventory: The skill generates markdown reports and documentation recommendations; it does not request tool permissions for arbitrary shell execution, system-level file writes, or outbound network requests.
  • Sanitization: The workflow does not specify validation or sanitization routines for instructions that may be embedded within the analyzed repository files.
Audit Metadata
Risk Level
SAFE
Analyzed
Jul 30, 2026, 10:35 AM
Security Audit — agent-trust-hub — vision-alignment-foresight