bug: prune thresholds are too aggressive for 1M+ context models

[okuyam2y]

Description

The prune function in SessionCompaction uses hardcoded thresholds (PRUNE_PROTECT = 40_000 and PRUNE_MINIMUM = 20_000 tokens) to decide when to clear old tool call outputs.

These values were reasonable when most models had 200K context (40K = 20%), but with Claude Opus 4.6, GPT-5.4, and Gemini 2.5 Pro all at 1M+ context, 40K is only 4%. Tool results from just a few turns ago get pruned even though context usage is well under the limit.

Steps to reproduce

1. Use a 1M context model (Claude Opus, GPT-5.4, etc.)
2. Have a multi-turn conversation with tool calls (reading/editing files)
3. After several turns, observe that tool outputs from earlier turns are pruned prematurely
4. The model loses access to recent work results

Expected behavior

Prune thresholds should scale with model context size so that larger-context models retain more tool output history.

[github-actions]

This issue might be a duplicate of existing issues. Please check:

• Auto-compaction triggers at ~200k tokens instead of model's actual context window (1M) #15871: Same root problem — auto-compaction/pruning uses hardcoded thresholds that don't account for 1M context windows, causing premature context truncation with Claude Opus 4.6 and similar models
• How to enable 1M context in OpenCode for GPT-5.4? Compaction triggers at 272k, not 1M #16308: Same symptom reported for GPT-5.4 — compaction triggers far below the model's actual 1M context window
• Feature: Configurable compaction threshold and model (global + per-model) #11930: Feature request for configurable compaction thresholds per model, which overlaps with the fix proposed here

[okuyam2y]

Related but different from #15871. That issue is about the compaction trigger threshold (when to start compacting), while this one is about the prune thresholds (how much tool output to protect from clearing). Both stem from hardcoded values that don't scale to 1M contexts, but they affect different code paths — overflow.ts vs compaction.ts:prune().

PR #21209 addresses only the prune side.

[okuyam2y]

Also related to #16308 (1M context but compaction/prune triggers at ~272K) — the hardcoded prune thresholds are part of that problem. #11930 requests making thresholds configurable; this PR takes the simpler approach of auto-scaling to context size.