koda-learn
Warn
Audited by Snyk on Jul 21, 2026
Risk Level: MEDIUM
Full Analysis
MEDIUM W011: Third-party content exposure detected (indirect prompt injection risk).
- Third-party content exposure detected (high risk: 0.85). Да: workflow «Установка» явно запрашивает и сохраняет в
.koda/skills/<name>/SKILL.mdсырой текст/skillMd/contentиз внешних веб-источников (public web catalogs и raw URLs), который затем станет LLM-читаемым содержимым для последующих шагов агента.
MEDIUM W012: Unverifiable external dependency detected (runtime URL that controls agent).
- Potentially malicious external URL detected (high risk: 0.90). The skill explicitly fetches and saves remote skill content at runtime (which can contain agent instructions) from URLs such as https://raw.githubusercontent.com/XCode-NLP/KodaSkills/refs/heads/main/metadata.json, https://agentskill.sh/api/agent/skills//install?owner=, and https://neuraldeep.ru/api/skills/readme?skillId=, so remote content can directly control prompts.
Issues (2)
W011
MEDIUMThird-party content exposure detected (indirect prompt injection risk).
W012
MEDIUMUnverifiable external dependency detected (runtime URL that controls agent).
Audit Metadata