Don Cheli: Context Health Monitor

Instructions

1. Estimate the current context window usage based on:

• Files loaded in the current session
• Conversation length (turns × average tokens)
• Any large outputs already in context

1. Report usage as a percentage of the model's context limit
2. Apply thresholds from the global rules:

• > 50K tokens in a result → recommend isolating in subagent
• > 200K session tokens → recommend summarizing and compressing context

1. Identify the highest-cost items in the current context:

• Large file reads
• Repeated content
• Verbose tool outputs that could be summarized

1. Recommend specific actions:

• Which files can be evicted (already processed)
• Which outputs should be summarized
• Whether a subagent should handle the next task

1. If context is healthy (< 40% used), confirm and suggest no action
2. Never truncate or discard context without explicit user confirmation

Output Format

## Context Health — 2026-03-28T16:00Z

### Usage Estimate
- Current session: ~85K tokens
- Model limit: 200K tokens
- Usage: 42% ✅ (healthy)

### Largest Items
1. src/payments/processor.ts full read — ~8K tokens (can evict — already analyzed)
2. Test run output from 14:32 — ~12K tokens (can summarize — report saved)
3. Conversation history — ~18K tokens (normal)

### Recommendations
- No immediate action required
- If you add 3+ more large files, consider /dc:destilar to compress session notes
- Next heavy task: use a subagent to isolate output

### Warning Threshold
Will warn again at 70% usage (~140K tokens)