ai-systems-coach

Pass

Audited by Gen Agent Trust Hub on Jul 17, 2026

Risk Level: SAFEPROMPT_INJECTION
Full Analysis
  • [PROMPT_INJECTION]: The skill instructions create a surface for indirect prompt injection by requiring the agent to ingest untrusted user data (stock and flow names) and output them verbatim in analysis and Mermaid diagrams.
  • Ingestion points: User-supplied structured descriptions of stocks, flows, and loops provided in the interaction flow defined in SKILL.md.
  • Boundary markers: None; the instructions do not mandate the use of delimiters or 'ignore' instructions when interpolating user-provided variable names.
  • Capability inventory: The skill is configured with Bash, Write, Read, and AskUserQuestion tools, allowing for file system interaction and command execution.
  • Sanitization: None; the instructions explicitly state that 'Variable names inside the diagram always preserve the user's original wording verbatim,' which prevents the sanitization of potentially malicious strings or command-like syntax.
Audit Metadata
Risk Level
SAFE
Analyzed
Jul 17, 2026, 02:04 PM
Security Audit — agent-trust-hub — ai-systems-coach