菜单
Use when interpreting Clawdad Watchtower review cards, especially soft versus hard findings in enforce mode.
用 Codex 或 Claude 帮你安装 复制这段 Prompt，粘贴到 Codex、Claude 或其他助手里，让它检查 Skill 页面并帮你完成安装。
Use when supervising a Clawdad delegate lane, converting soft review signals into next prompts and preserving hard safety stops.
Use when acting as a Clawdad delegate session for repo work, including bounded implementation, validation, and hard-stop reporting.
Use when Clawdad failures, repeated failed messages, Watchtower pauses, session import issues, or delegate stalls need root-cause triage.
Use when cutting, publishing, installing, and verifying a Clawdad release across npm, git tags, GitHub releases, and the local service.
Use when diagnosing Clawdad/Codex project session IDs, imported sessions, stale active pointers, quarantines, and delegate lane bindings.
Use when supervising a Clawdad delegate lane, converting soft review signals into next prompts and preserving hard safety stops.
基于 SOC 职业分类
| name | clawdad-watchtower-review |
| description | Use when interpreting Clawdad Watchtower review cards, especially soft versus hard findings in enforce mode. |
Review Watchtower output as policy signal.
Classify findings this way:
For hard stops, preserve the pause and explain the gate. For soft findings, produce the next corrective prompt for the delegate.
