skill-creator

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

• Third-party content exposure detected (high risk: 0.90). The SKILL.md installation workflow (Step 1 — "Obtain skill content") explicitly instructs the agent to fetch content from URLs (via curl or web_fetch) and then locate/load the downloaded SKILL.md and any references/scripts into context, so arbitrary public/untrusted web content or archives can be read and can materially influence the agent's actions.

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

• Potentially malicious external URL detected (high risk: 0.90). The skill explicitly performs runtime fetches of SKILL.md or archives via curl/web_fetch when installing a skill from a URL (e.g., any URL such as https://.../SKILL.md or an archive URL), and the fetched SKILL.md is loaded into the agent's context and therefore directly controls prompts/instructions as a required dependency for installation.