docs-reading

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

• Third-party content exposure detected (high risk: 0.80). The skill's instructions (SKILL.md and scripts/README.md) explicitly ingest public third-party docs (e.g., cloning and parsing https://github.com/godotengine/godot.git or arbitrary Markdown via scripts/parse_docs.py) into a local SQLite DB that the agent queries with agent-docs-search at runtime, so untrusted external content can be read and materially influence the agent's decisions.