cawplan-internal-qa-coding-humaninputs-test

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

• Third-party content exposure detected (medium risk: 0.30). The skill reads human input logs and coding inputs via CLI commands (cawplan session human-input-logs and product-human-input-logs) which ingest user-submitted content from team members interacting with the coding platform.