Overview

nancy-send-message lets your agent report blockers, progress, and review requests to the orchestrator during autonomous execution. It uses the ralph msg <type> \"<message>\" format to keep updates consistent and actionable. This standardization helps teams act quickly when issues arise or reviews are needed.

How This Skill Works

The skill formats a message with a type (blocker, progress, or review-request) and sends it to the orchestrator via the ralph msg command. The orchestrator uses these updates to guide decisions and monitor autonomous execution.

When to Use It

• When a blocker prevents progress (e.g., missing credentials)
• When you need to report notable progress to stakeholders
• When a feature is complete and a review is required
• During autonomous runs to communicate current status
• When you need to escalate or request guidance from the orchestrator

Quick Start

1. Step 1: Determine the appropriate type (blocker, progress, or review-request) and craft your message.
2. Step 2: Send the message using: ralph msg <type> "<message>" (examples above).
3. Step 3: Verify delivery in the orchestrator logs or UI.

Best Practices

• Use the exact types: blocker, progress, or review-request
• Keep each message concise (one sentence preferably)
• Avoid exposing secrets or credentials in the message
• Format messages as: ralph msg <type> \"<message>\"
• Test with a dry run or sample messages before enabling automation

Example Use Cases

• ralph msg blocker "Cannot proceed: missing API credentials"
• ralph msg progress "Completed database migration"
• ralph msg review-request "Feature complete, ready for verification"
• ralph msg progress "Cache warmed for region us-east"
• ralph msg blocker "Rate limit reached; retry later"

Frequently Asked Questions