koda-helper

Pass

Audited by Gen Agent Trust Hub on Jul 21, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADSPROMPT_INJECTION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill references multiple official resources, including documentation at docs.kodacode.ru, GitHub repositories under the XCode-NLP organization, and official plugin marketplaces. It also provides template configurations for MCP servers that suggest the use of npx. These references are part of the intended functionality for a technical support skill and point to vendor-controlled or well-known service domains.
  • [PROMPT_INJECTION]: The instructions include a mandate to fetch content from external URLs using tools like fetch_url_content when context is missing. While this establishes a surface for indirect prompt injection (ingesting untrusted data), the skill specifically directs the agent toward official and trusted documentation resources. There are no explicit boundary markers or sanitization instructions for this ingested data, but the behavior is restricted to the skill's primary purpose of providing product support.
Audit Metadata
Risk Level
SAFE
Analyzed
Jul 21, 2026, 08:08 AM
Security Audit — agent-trust-hub — koda-helper