blueprint-derive-adr

Pass

Audited by Gen Agent Trust Hub on May 14, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: The skill operates on local project files to generate Architecture Decision Records (ADRs). All command-line operations (ls, mkdir, grep, wc) are restricted to local file management and information gathering.
  • [INDIRECT_PROMPT_INJECTION]: The skill analyzes existing project content, which represents a surface for indirect prompt injection as it ingests untrusted data into the agent context.
  • Ingestion points: Codebase structure, dependencies, and documentation are analyzed using the Explore subagent and file-reading tools (SKILL.md).
  • Boundary markers: Absent; the prompt for the subagent does not define delimiters to distinguish between code and instructions.
  • Capability inventory: The skill possesses file-writing capabilities (Write tool) and local shell execution (Bash tool) (SKILL.md).
  • Sanitization: No explicit sanitization or filtering of codebase content is performed before processing.
Audit Metadata
Risk Level
SAFE
Analyzed
May 14, 2026, 04:17 PM
Security Audit — agent-trust-hub — blueprint-derive-adr