urithiru

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

• Third-party content exposure detected (high risk: 0.95). At runtime, the skill’s router uses the user-provided --query text (outsider-authored free text) as the --query/--query-file input to runner.py, which is then embedded/sent into the lane models’ LLM context; this is direct prompt-injection exposure from arbitrary user text.

MEDIUM W021: Hidden or invisible Unicode characters detected (potential obfuscation or prompt injection).

• Hidden Unicode characters detected (1 type(s) found)