new-milestone

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

• Third-party content exposure detected (high risk: 0.75). At runtime the skill reads the existing tunan:project GitHub issue body via gh issue view <N> --json body --jq .body, and that issue content is outsider-authored free text (any non-operating-user edits to the issue), which is then ingested into the agent’s LLM context for parsing/summarization.