opencode-session-debugging

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

• Third-party content exposure detected (medium risk: 0.30). The skill ingests outsider-authored free text at runtime only via user-provided session identifiers/titles/symptoms that are used to read local transcript files and logs (e.g., ~/.claude/transcripts/*.jsonl), which may contain prompts originating from outsiders; therefore the required workflow can end up processing that free text.