paperclip

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

• Third-party content exposure detected (high risk: 0.90). The skill's Company Skills Workflow explicitly instructs the agent to import and inspect skills from external sources such as skills.sh URLs or arbitrary GitHub URLs (references/company-skills.md) and to read installed SKILL.md files, which are untrusted third‑party/user-provided content that can change agent behavior.

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

• Potentially malicious external URL detected (high risk: 0.90). The skill explicitly supports importing skills at runtime from user-provided URLs (e.g. https://skills.sh/google-labs-code/stitch-skills/design-md and https://github.com/vercel-labs/agent-browser), which the system fetches and installs as SKILL.md/instruction content that can directly control agent behavior, so these external URLs are runtime dependencies that can change agent prompts/instructions.