firewall-review
Pass
Audited by Gen Agent Trust Hub on Jul 20, 2026
Risk Level: SAFEPROMPT_INJECTION
Full Analysis
- [PROMPT_INJECTION]: The skill exhibits a surface for indirect prompt injection. It is designed to ingest and process external firewall configuration files (e.g., iptables, Terraform, Kubernetes YAML) which could contain attacker-controlled comments or metadata.
- Ingestion points: Target files and directories discovered via Glob and Grep patterns (Step 1).
- Boundary markers: The skill includes a dedicated 'Prompt Injection Safety Notice' section instructing the agent to ignore comments and treat all configuration content as untrusted data.
- Capability inventory: Access is restricted to 'Read', 'Grep', and 'Glob' tools. The skill does not request network access or the ability to execute shell commands or arbitrary code.
- Sanitization: The instructions explicitly warn the agent not to interpret configuration comments as instructions or evaluate expressions found within rule descriptions.
Audit Metadata