capture-desktop

HIGH W007: Insecure credential handling detected in skill instructions.

• Insecure credential handling detected (high risk: 1.00). The skill instructs the agent to accept/paste a full desktop transcript, stage and read a verbatim transcript, and "show the exact content you will write" and propose exact edits—forcing the LLM to reproduce transcript text (which may include API keys/passwords) verbatim in its outputs and writes.

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

• Third-party content exposure detected (medium risk: 0.30). SKILL.md specifies that the agent reads the staged transcript produced from Ane’s pasted/locally provided Desktop conversation (Step 2–3), so outsider-authored free text is ingested directly into runtime LLM processing and then routed into outputs.