knowhow-distill

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

• Third-party content exposure detected (high risk: 0.75). Runtime path: the workflow reads .knowhow/inbox/*.md (outsider-authored free text from other people’s messages/comments) and also reads .knowhow/schema-signals.md lines, then “Ghi nhận toàn bộ vào context” and “Đọc nội dung từng file” before analysis—so any outsider text placed into inbox can be ingested into the agent’s LLM context.

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

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