gemini-docs-skill

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

• Third-party content exposure detected (high risk: 0.85). Outsider source: the skill’s runtime “update gemini documentation” path calls scripts/update-docs.sh, which uses _shared/doc-builder.py with docs-config.yaml to fetch public web pages from https://ai.google.dev/... (including llms.txt and listed *.md.txt URLs) and then ingests their readable text into the generated references/*.md / full-docs.txt that the agent later reads for LLM context.