[ranxianglei]

Closes #421

Problem

V2 native compaction projects the checkpoint as an assistant carrying the compaction request's token usage verbatim (lib/v2/projection/shared.ts makeAssistantInfo). cacheSystemPromptTokens used that assistant as its calibration anchor, permanently storing the request size as "system + tool schemas" overhead. Repro from the issue: input 653137 → stored 653136 phantom; on a 400000-token model the hard-guard budget becomes −285904, producing negative-budget log spam and unnecessary tool clearing while real provider usage stays far smaller.

Fix

• lib/token-utils.ts — new calibrateSystemOverhead(state, messages): never anchors on info.summary === true assistants or assistants created before state.lastCompaction; subtracts the full wire-visible pre-anchor prefix with BPE-exact counts (excluding host-ignored user parts and ACP-owned notices). estimateSystemPromptTokens(messages, lastCompaction = 0) delegates to it.
• lib/ui/utils.ts — cacheSystemPromptTokens(state, messages, measuredSystemTokens?) merges the heuristic residual with a measured current-wire floor via max(), recording provenance in new transient field SessionState.systemPromptTokensSource (heuristic | measured). The [FIX estimateSystemPromptTokens inconsistent between nudge and acp_status after compression #255] stable-cache early return is preserved.
• lib/v2/context.ts — measures the actual outgoing system parts (Σ countTokens(event.system[].text) + rendered ACP system prompt) each transaction and passes it as the floor. V1 callers omit it → pure heuristic, unchanged.
• Invalidation — resetOnCompaction (lib/state/utils.ts) and the mid-session model-switch block (lib/messages/transform.ts) now clear the cached estimate, so first-request-after-compaction and after reload recalibrate cleanly instead of inheriting a stale/phantom value.
• lib/compress/status.ts — /acp context breakdown labels the system segment [measured] / [estimated], distinguishing actual wire accounting from heuristics (issue acceptance criterion).

No persisted-state format change (systemPromptTokens* were already transient), no config change, all API extensions are optional trailing parameters.

Verification

• New regression suite tests/v2-compaction-system-overhead.test.ts — 14/14 pass: the exact V2: compaction usage becomes phantom system overhead and negative context budgets #421 repro (fresh state + reload both store no phantom), post-compaction anchor calibrates to real wire overhead with nonnegative budget at the reported 400k limit, measured-floor/provenance behavior, invalidation paths, and estimateSystemPromptTokens inconsistent between nudge and acp_status after compression #255 compatibility (plain-text sessions calibrate byte-for-byte as before).
• Bug-catch verified: with lib/ reverted to pre-fix code, the original cacheSystemPromptTokens on the repro input stores exactly 653136; fixed code stores undefined.
• Full suite: 1424/1425 (sole failure = tests/soft-block.test.ts module-load EACCES /tmp — sandbox mounts /tmp read-only, pre-existing environmental issue, unrelated).
• npm run typecheck clean, npm run build green, touched files prettier-clean.
• Devlog: devlog/2026-09-17_v2-compaction-phantom-system/{REQ,DESIGN,WORKLOG}.md.

中文摘要:修复了 V2 原生压缩后校准锚点误用 compaction 请求用量、把 653136 tokens 当成永久 system 开销导致硬护栏预算为负(-285904)的问题——排除 summary/过期锚点、按当前实际出站 system 部件做测量下限、在压缩与模型切换时失效缓存,并在 /acp context 中标注 measured/estimated 来源。

[ranxianglei]

🤖 Powered by ework · qwen3.8-27b

[bot] 🏷 收到，开始核验这个 PR（不重做，按 [PR] 流程：拉分支核对代码 + 确认 CI + 报告结论）。已确认 PR 基于 2026-09-17_v2-base（V2 迁移分支），head 为单提交 64f31509，GitHub 侧 mergeable/clean。正在本地跑新回归套件和全量测试，稍后给出完整结论。

[ranxianglei]

🤖 Powered by ework · qwen3.8-27b

核验完成（按 [PR] 流程：拉分支独立验证，未重做）。结论：修复正确、可合并，另补了一处漏网的小缺口（已直接修进 PR 分支）。

独立验证结果（全部通过）

• npm run typecheck 干净；npm run build 成功；涉及文件 prettier 干净。
• 新回归套件 tests/v2-compaction-system-overhead.test.ts：14/14 通过。
• 全量测试：1424/1425 通过——唯一失败是 tests/soft-block.test.ts 模块加载时 EACCES mkdir '/tmp/...'，本沙箱 /tmp 只读所致，属既有环境问题，该文件未被本 PR 触碰。
• 独立 bug-catch：在纯修复前代码上用 V2: compaction usage becomes phantom system overhead and negative context budgets #421 的原始复现输入跑旧 cacheSystemPromptTokens，存入值恰为 653136（即幻影开销）；修复后代码存 undefined。确认修复打中根因。
• 逐行审过全部 diff：锚点排除规则（summary===true / 早于 lastCompaction 创建）、BPE 精确前缀扣除（排除 host-ignored user parts 与 ACP-owned notices）、测量下限 max() 合并 + provenance 字段、V2-only 测量接线（V1 调用方不变）、压缩/模型切换两处失效、/acp context 的 [measured]/[estimated] 标注——均正确一致。无持久化状态格式变更（两个新字段确认未进 persistence 层）、无配置变更、未动 version，devlog REQ/DESIGN/WORKLOG 齐全。

审查中发现并已直接修复的一处缺口（commit e833eff3，已推回 PR 分支）
estimateContextComposition 的 live fallback（lib/messages/inject/utils.ts）调用估算器时没传 lastCompaction——resetOnCompaction 失效缓存后，压缩前的 assistant 会重新成为校准锚点，把过期开销重新带进 nudge/context-usage 计算。PR 已在同类的 lib/compress/status.ts fallback 里加了这层防护，唯独漏了这条路径。已补传 state?.lastCompaction ?? 0 并新增回归测试 #15（对修复前代码验证过会失败：systemTokens 取到 STALE 而非 WIRE；修复后通过）。补测后：新套件 15/15，全量 1425/1426（唯一失败仍为上述环境性 soft-block），typecheck/prettier 干净。WORKLOG 已追加记录。

给 owner 的两点提示

1. CI 不会在这个 PR 上跑：所有 workflow（ci.yml / pr-checks.yml / pr-artifact.yml / release.yml）都只触发 branches: [master]，而本 PR base 是 2026-09-17_v2-base。因此该分支没有任何 CI 检查位，以上本地验证即为把关依据。若希望 V2 系列有 CI 覆盖，可考虑把 V2 base 分支加进 workflow triggers。
2. soft-block 那条失败在 CI（/tmp 可写）里会通过。

合并请人工操作（我无权 merge）。

中文摘要：核验了 #421 幻影 system 开销修复的正确性（旧代码复现存入 653136、新代码不存），并直接修复了 live fallback 漏传 lastCompaction 的一处同类缺口（附回归测试，已推回 PR 分支），typecheck/build/全量测试均过（唯一失败为沙箱 /tmp 只读的环境问题）；另注意此 PR base 非 master，CI 不会运行；可以合并。