rovo-docs

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

• Third-party content exposure detected (high risk: 0.85). SKILL.md instructs the agent to use WebFetch at runtime to load arbitrary Rovo documentation pages from support.atlassian.com (and potentially community/status URLs), and the fetched page text is then extracted and fed into the LLM context—an outsider-authored public web content path.

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

• Potentially malicious external URL detected (high risk: 0.90). The skill explicitly instructs the agent to perform WebFetch at runtime to load and inject external documentation into the model context (e.g., https://support.atlassian.com/rovo/docs/work-with-rovo-dev/), so fetched remote content can directly influence prompts/outputs.