dx-alerts
Lightweight “news wire” implemented **via Agent Mail threads**, not a separate bespoke system.
When & Why to Use This Skill
The dx-alerts Claude skill is a lightweight, high-efficiency 'news wire' designed to streamline communication between AI agents using Agent Mail threads. It solves the common 'pull' limitation of agents by pushing critical, time-sensitive developer experience (DX) alerts—ranging from blockers to low-priority updates—directly into the agent's context. This ensures that tools like dx-doctor can surface relevant repository status updates across prime-radiant-ai, affordabot, and llm-common without blocking workflows.
Use Cases
- Critical Blocker Notification: Automatically pushing [blocker] tags to agents when a core repository build fails, ensuring immediate awareness and response.
- Cross-Agent Coordination: Using the dx-alerts thread to synchronize high-priority task status and system health across different specialized agents.
- Repository Health Tracking: Providing agents with a filtered, severity-based feed of maintenance notes and updates for specific codebases like llm-common or affordabot.
| name | dx-alerts |
|---|---|
| description | Lightweight “news wire” implemented **via Agent Mail threads**, not a separate bespoke system. |
dx-alerts
Lightweight “news wire” implemented via Agent Mail threads, not a separate bespoke system.
Pattern:
- Publish alerts into an Agent Mail thread id:
dx-alerts - Each alert message should start with a tag:
[DX-ALERT][blocker] ...[DX-ALERT][high] ...[DX-ALERT][medium] ...[DX-ALERT][low] ...
- Include repo scope in the body (prime-radiant-ai / affordabot / llm-common) and an expiry if relevant.
Consumption:
dx-doctorshould optionally surface the latest Ndx-alertsmessages (best-effort; never block).
Why:
- Avoids “agents don’t pull agent-skills” problems by pushing time-sensitive info into a mailbox.