aif

MEDIUM W011: Third-party content exposure detected (indirect prompt injection risk).

• Third-party content exposure detected (high risk: 0.90). SKILL.md explicitly instructs the agent to search/install external skills from skills.sh and to use "Learn Mode" by passing arbitrary documentation URLs to /aif-skill-generator so the agent will fetch and study public web pages and docs (see "Skill Acquisition Strategy" and "Learn Mode" in SKILL.md), which lets untrusted third‑party content be ingested and influence generated skills and subsequent actions.

MEDIUM W012: Unverifiable external dependency detected (runtime URL that controls agent).

• Potentially malicious external URL detected (high risk: 0.90). The skill explicitly installs and runs external skills from https://skills.sh (via npx) and accepts arbitrary documentation URLs passed to /aif-skill-generator in "Learn Mode", which are fetched at runtime and can directly control generated prompts or cause remote code to be executed.