skills/vercel-labs/ai/adr-skill/Gen Agent Trust Hub

adr-skill

Pass

Audited by Gen Agent Trust Hub on May 9, 2026

Risk Level: SAFE
Full Analysis
  • [Architecture and Workflow]: The skill implements a robust, four-phase workflow (Scan, Capture, Draft, Review) designed to maintain high-quality architectural documentation. These instructions are focused on repository maintenance and do not contain patterns associated with prompt injection or malicious behavior.
  • [Local Utility Scripts]: The skill includes several JavaScript utilities (bootstrap_adr.js, new_adr.js, set_adr_status.js) to automate file management. These scripts use standard Node.js modules (fs, path) to interact with the local filesystem. Analysis confirms they do not perform network operations, download external content, or access sensitive system directories.
  • [Contextual Repository Scanning]: The skill guides the agent to scan common project files (such as package.json or go.mod) to gather relevant context for architectural decisions. This is a standard and expected operation for a documentation-focused skill and does not involve the exposure of credentials or sensitive data.
  • [Templates and References]: The provided templates and reference materials follow industry-standard formats (including MADR 4.0). They are designed to assist the agent in drafting clear, executable specifications and do not contain any obfuscated content or suspicious links.
Audit Metadata
Risk Level
SAFE
Analyzed
May 9, 2026, 05:35 AM