# Sliding Window Attention Audit

## Description
Audits sliding window context buffers and attention sink token retention to bound KV memory in continuous streaming.

## Category
`architecture`

## Tags
`sliding-window`, `attention-sink`, `streaming`, `kv-cache`, `memory`

## When to Use
Audit continuous streaming deployments to ensure KV memory remains strictly bounded without perplexity divergence.

## Operational Limitations
Evaluates window buffer size and attention sink retention parameters.

## Verification Script
Run the deterministic verification checker:
```bash
python3 scripts/check.py examples/synthetic.json
```
Exit codes:
- `0`: Check passed and thresholds satisfied.
- `1`: Threshold breach or operational violation detected.
- `2`: Schema error or invalid usage.
