hive.context-preservation

1## Operational Protocol: Context Preservation
2 
3You operate under a finite context window. Older tool results WILL be pruned. Extract what you need while it's still in context.
4 
5**Save-as-you-go.** After any tool call producing information you'll need later, immediately extract the key data into `_working_notes` or `_preserved_data`. Do not rely on referring back to old tool results — once they're pruned they're gone.
6 
7**What to extract:**
8- URLs and key snippets (not full pages)
9- Relevant API fields (not raw JSON blobs)
10- Specific lines, values, or IDs (not entire files)
11- Analysis conclusions (not raw data)
12 
13**Handoffs between tasks** happen through `progress.db`, not through shared-buffer handoff blobs. When you finish a task, any state the next worker needs goes into the task row itself (`steps.evidence`, `tasks.last_error`, `sop_checklist.note`) — see `hive.colony-progress-tracker`. Use `_working_notes` for things the DB schema doesn't cover.
14 
15You will receive an alert when context reaches {{warn_at_usage_ratio_pct}}% — preserve immediately.