[plan] Address workflow reliability and backlog hygiene issues

[github-actions]

Overview

This tracking issue addresses critical findings from the DeepReport Intelligence Briefing (Discussion #6855), focusing on improving workflow reliability, reducing noise, and maintaining backlog hygiene.

Source: Discussion #6855

Problem Statement

Recent analysis revealed three key concerns:

1. Workflow reliability: 3 of last 10 runs failed/cancelled, with successful runs showing high error/warning counts
2. Token cost spike: ~4x increase to 1.71M tokens suggests verbose prompts or inefficient analyses
3. Backlog growth: +12 issues week-over-week with label skew (124 plan/ai-generated each, 27 unlabeled)

Planned Tasks

This work is broken into 5 focused sub-issues:

1. Reduce workflow log noise - Clean up error/warning spam in successful runs for better signal
2. Optimize prompt token usage - Trim verbosity in high-cost runs to reduce token spend
3. Triage unlabeled issues - Process 27 unlabeled issues to improve backlog clarity
4. Investigate workflow failures - Analyze root causes of recent 3/10 failure rate
5. Review label distribution - Address skew toward plan/ai-generated labels

Success Metrics

• Workflow failure rate below 10% (from 30%)
• Token usage reduced by 50% without quality loss
• Zero unlabeled open issues (from 6)
• Clear ownership/priority signals on backlog items

AI generated by Plan Command for discussion #6855