custom-domains

Pass

Audited by Gen Agent Trust Hub on Nov 20, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
  • [SAFE]: The skill provides legitimate documentation for the Internet Computer's custom domain service. All referenced infrastructure (icp.net, icp-api.io), CLI tools (icp), and software libraries (@dfinity/static-site, @icp-sdk/core) are official resources provided by the vendor, dfinity.
  • [INDIRECT_PROMPT_INJECTION]: The skill includes workflows where user-supplied inputs (domain names and canister IDs) are used to construct shell commands for validation and registration.
  • Ingestion points: Untrusted user data is ingested via the CUSTOM_DOMAIN and CANISTER_ID placeholders used throughout the registration and troubleshooting steps in SKILL.md.
  • Boundary markers: The instructions use double quotes around placeholders in shell templates to provide basic delimitation.
  • Capability inventory: The skill utilizes curl for interacting with the registration REST API and the icp CLI for making canister calls (SKILL.md).
  • Sanitization: The instructions rely on the user or agent to correctly interpolate the variables; no explicit sanitization or validation logic is defined within the skill itself. As this is the primary purpose of the skill, the risk is considered low.
Audit Metadata
Risk Level
SAFE
Analyzed
Nov 20, 2026, 09:53 PM