acomm-receive

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

• Third-party content exposure detected (high risk: 1.00). The SKILL.md shows the skill runs "acomm --receive" to read messages from third-party channels (Discord/Slack/ntfy) and outputs the raw user-generated message to stdout for use in subsequent branching (see the "手動確認フロー" and "注意事項" examples), so untrusted external content can directly influence agent behavior.