segmentation

Pass

Audited by Gen Agent Trust Hub on May 8, 2026

Risk Level: SAFEPROMPT_INJECTION
Full Analysis
  • [PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection because it processes untrusted external data in the form of network configuration and infrastructure-as-code files.
  • Ingestion points: The process involves searching for and reading files such as Terraform (.tf) definitions, Kubernetes NetworkPolicies, and firewall configurations using the Glob and Grep tools.
  • Boundary markers: The skill contains a 'Prompt Injection Safety Notice' which explicitly warns the agent not to interpret configuration comments or resource tags as instructions.
  • Capability inventory: The skill's capabilities are limited to local file system discovery and reading (Read, Grep, Glob) with no access to shell execution or network tools.
  • Sanitization: The instructions specifically direct the agent to treat all configuration content as untrusted data and to ignore any text that appears to be a prompt or command.
Audit Metadata
Risk Level
SAFE
Analyzed
May 8, 2026, 12:28 AM
Security Audit — agent-trust-hub — segmentation