configuring-ip-allowlists

Pass

Audited by Gen Agent Trust Hub on Mar 23, 2026

Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill uses the ccloud and cockroach CLI tools to manage network security settings. These commands are necessary for the skill's intended purpose of hardening cluster access and are standard administrative tools for this environment.
  • [EXTERNAL_DOWNLOADS]: Uses curl to retrieve the user's public IP address from checkip.amazonaws.com, a well-known service. This is a legitimate utility step to ensure the administrator's current IP is whitelisted before removing broader access, preventing accidental lockout.
  • [PROMPT_INJECTION]: The skill identifies an indirect prompt injection surface where user-supplied inputs such as cluster identifiers and CIDR ranges are interpolated into shell commands.
  • Ingestion points: User-provided values for cluster names, IDs, and network CIDR blocks within SKILL.md.
  • Boundary markers: None; uses standard markdown formatting for command blocks.
  • Capability inventory: The skill utilizes ccloud, cockroach, and curl through shell execution.
  • Sanitization: No explicit sanitization or validation of the input strings is described in the documentation text.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 23, 2026, 10:27 PM
Security Audit — agent-trust-hub — configuring-ip-allowlists