improve-architecture

Pass

Audited by Gen Agent Trust Hub on Sep 14, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted data from project documentation and source code to identify architectural issues and design debt.\n
  • Ingestion points: The skill reads domain context from CONTEXT.md, historical decisions from docs/adr/, product requirements from docs/prd/*.md, and technical signals from the entire codebase.\n
  • Boundary markers: There are no explicit delimiters or instructions provided to the agent to distinguish between its internal instructions and the content of the files being analyzed.\n
  • Capability inventory: The skill's primary function is to perform read-only analysis and generate a text-based report; it lacks the capability to make network requests or autonomously write code changes, though it may invoke git log for change frequency analysis.\n
  • Sanitization: No specific validation or sanitization routines are defined for the data extracted from the analyzed files before it is processed by the agent.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 14, 2026, 06:21 AM
Security Audit — agent-trust-hub — improve-architecture