continuous-learning-v2

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

• Third-party content exposure detected (high risk: 0.90). The CLI explicitly supports importing instincts from arbitrary http/https URLs (scripts/instinct-cli.py — cmd_import uses urllib.request.urlopen to fetch remote content) and then parses/writes those imported instincts into project/global instincts that the agent later loads and acts on, meaning untrusted third‑party content can directly influence agent behavior.