cowork-export

HIGH W007: Insecure credential handling detected in skill instructions.

• Insecure credential handling detected (high risk: 1.00). The skill instructs the agent to read and export Qdrant memory, inline file contents, and git state into a clipboard-ready briefing (and copy it to the clipboard), which will verbatim include any API keys, tokens, or passwords present in those sources unless explicitly redacted.

MEDIUM W012: Unverifiable external dependency detected (runtime URL that controls agent).

• Potentially malicious external URL detected (high risk: 0.80). The script makes a runtime HTTP POST to QDRANT_URL (e.g., http://localhost:6333/collections/agent_memory/points/scroll) and directly injects the returned memory "content" into the briefing that becomes the remote agent prompt, so remote Qdrant-hosted data can control the agent's instructions.