segmentation

Pass

Audited by Gen Agent Trust Hub on Aug 26, 2026

Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill uses dynamic context injection !$ARGUMENTS in the 'When to Use' section to customize the review focus based on user input. This is a common pattern in development-oriented skills and is used here for scoping purposes.
  • [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted infrastructure-as-code (IaC) files and network configurations. It explicitly includes a safety notice to mitigate this risk by instructing the agent to ignore instructions embedded in configuration comments or tags.
  • Ingestion points: Reads files matching various patterns like **/*.tf, **/NetworkPolicy*, and **/firewall-rule* (SKILL.md, Step 1).
  • Boundary markers: Includes a 'Prompt Injection Safety Notice' section advising the agent to treat configuration content as untrusted data.
  • Capability inventory: Uses Read, Grep, and Glob tools to analyze files.
  • Sanitization: Relies on instructional guardrails to prevent interpretation of data as commands.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 26, 2026, 01:48 PM
Security Audit — agent-trust-hub — segmentation