@danceiny/gotry 0.0.1-rc.15 → 0.0.1-rc.17

package/README.md

@@ -12,7 +12,7 @@
 | | |
 |---|---|
 | **Version** | `v0.0.1-rc.13+` (npm `latest`; [release notes](docs/release-notes.md)) |
-| **Runtime** | DeepSeek Harness **0.1.2-alpha.1** (vendored; [upstream](https://github.com/deepseek-ai/DeepSeek-Harness)) · Z3 WASM · Cordis |
+| **Runtime** | DeepSeek Harness **0.1.2-alpha.3** (root-pinned; [upstream](https://github.com/deepseek-ai/DeepSeek-Harness)) · Z3 WASM · Cordis |
 | **License** | **MIT** ([LICENSE](LICENSE)) |
 | **Docs** | English (this file) · [简体中文 README](README.zh-CN.md) · deep engineering docs are Chinese-first ([docs/architecture.md](docs/architecture.md)) |
 
@@ -23,16 +23,25 @@
 ```bash
 npx @danceiny/gotry web
 # First run creates .env: LLM_API_KEY=<DeepSeek key, or OpenAI-compatible key>
+#   Not using DeepSeek directly? Also set LLM_BASE_URL=<your endpoint, usually ending in /v1>
+#   Need a specific model (relays often do)? Set LLM_MODEL=<model name> — it drives the
+#   dsh chat face (web/headless) and overrides any model picked in the dsh web settings.
 # → open http://127.0.0.1:3080 and chat: "I want three relaxing days in Dali"
 ```
 
+> **Switching models / providers?** The price table (`ts/data/llm-price-table.json`, schema `gotry_llm_price_table_v2`) is the single source of truth for `gotry_m3_nightly_run_v1.cost_usd`. Adding a new model or switching relay = update this file via PR (ADR-11, peak-conservative upper bound only)
+- . Unknown models **fail-closed** — no guessed prices. Drift monitor: `npx tsx ts/scripts/price-drift-watch.ts` (offline baseline diff;
+- `--fetch` for live official pages). Never auto-applies changes.
+
 | You want | Command |
 |---|---|
 | 🖥️ Conversational planner (recommended) | `npx @danceiny/gotry web` → chat UI on :3080 |
 | 🤖 Scripted / one-shot answer | `npx @danceiny/gotry "Two recovery days from Shenzhen, budget 3000"` |
 | 🛠️ Developer: run from source | see [source install](#%EF%B8%8F-developer-source-install) below |
 
-- Requires Node 22+ and one LLM API key. Zero-config startup — the dsh runtime is mounted automatically via a cordis patch.
+- Requires Node 22.15+ and one LLM API key. Any OpenAI-compatible endpoint (MiniMax / relays / self-hosted gateways) works too — add `LLM_BASE_URL` to `.env` (usually ends with `/v1`, e.g. `https://api.minimax.io/v1`) and requests follow it instead of the DeepSeek default. To pin the model,
+  - set `LLM_MODEL` (e.g. `MiniMax-M2`): it applies to both the dsh chat face and the repo scripts, and beats the model selection persisted in the dsh web UI;
+  - unset, the dsh built-in default (`deepseek-v4-flash`) or your web-UI choice is used. Zero-config startup — the dsh runtime is mounted automatically via a cordis patch.
 
 ---
 
@@ -66,9 +75,10 @@ That's it — the dsh chat UI on `:3080` with the GoTry persona mounted. First c
 
 ```bash
 git clone https://github.com/Danceiny/gotry && cd gotry
-cd ts/dsh-runtime && pnpm install && cd ../..    # ① vendored dsh 0.1.2-alpha.1 (one-off)
-cp .env.example .env                              # ② set LLM_API_KEY
-./gotry web                                       # ③ in-repo entry, same UX
+npm ci && npm --prefix ts ci                      # ① install the pinned root/TS closure
+node scripts/build-dist.mjs                       # ② build the source checkout's JS runtime
+cp .env.example .env                              # ③ set LLM_API_KEY (+ LLM_BASE_URL if not on DeepSeek; + LLM_MODEL to pin the model)
+./gotry web                                       # ④ in-repo entry, same UX
 ```
 
 | Entry | Command | When |
@@ -76,9 +86,11 @@ cp .env.example .env                              # ② set LLM_API_KEY
 | dsh Web chat (recommended) | `./gotry web` | multi-turn planning with visualized reasoning → :3080 |
 | headless one-shot | `./gotry "one full task"` | scripts / CI / targeted debugging → stdout |
 
+The source entry and npm package both resolve one 216-package DSH `0.1.2-alpha.3` closure declared as exact direct dependencies. The manifest, package lock, and root pnpm importer must expose the same 216-name set; publish preverify rejects omissions, mixed versions, and ranges. Source checkout runs dsh from `ts/dsh-runtime/` in normal mode so `gotry-state/` continuity is preserved; benchmark opt-in and npm-package runs use the invocation directory for isolation. The old `ts/dsh-runtime/` vendor tree remains a non-benchmark legacy resolution fallback, not a promised runnable path or the recommended install path.
+
 ---
 
-## 🧰 20 tools
+## 🧰 21 tools
 
 | Group | Tool | What it does |
 |---|---|---|
@@ -95,6 +107,7 @@ cp .env.example .env                              # ② set LLM_API_KEY
 | | `gotry_wish_pool_add` / `gotry_wish_pool_list` | "next departure" wish pool + 0..1 conditional recall |
 | | `gotry_companion_save` · `gotry_trip_log` | companion profile / travel timeline |
 | **Artifacts** | `gotry_artifacts_list` / `gotry_artifacts_read` | Discover & view generated artifacts (async deliverables + working-dir markdown) as a line-numbered file view (read-only) |
+| **Factuality gate** | `gotry_fact_gate` | Pre-delivery gate for itinerary artifacts: every bookable claim (flight no./time/airport/price/policy) must trace to an exact-date tool result (hit AND miss recorded); unverifiable ⇒ blocked — never present as a verified plan |
 | **General external** | `gotry_web_search` · `gotry_video_subtitle` · `gotry_github_search` · `gotry_agent_reach` | web / subtitles / GitHub / all-channel external info (via Agent-Reach) |
 
 ---
@@ -108,7 +121,8 @@ The account session channel reads realtime hotel/flight data from **the user's o
 3. **Physically read-only.** A ReadGuard aborts all write requests at the network layer (ordering/payment is unreachable in transport); the agent never touches credentials or captchas — on a captcha it stops and hands control back.
 4. **Never hijacks your browser.** Retrieval/login always open their own dedicated tab; the login page is brought to front and stays with you; routine test runs never open browser windows.
 
-> One-time prerequisite: in your daily Chrome, open `chrome://inspect/#remote-debugging` and enable the switch (Chrome 144+). Without it the tools return `needs-attach` with instructions and spend nothing.
+> One-time prerequisite: install the [GoTry Session Bridge](https://chromewebstore.google.com/detail/gotry-session-bridge/oeajpiccmonococjcegddlooeeohlbgd) Chrome extension from the Chrome Web Store (one click, auto-updates) — done in the browser, not by gotry. Review-free GitHub-channel fallback is optional; for that variant see [`docs/extension-webstore-submission.md`](docs/extension-webstore-submission.md). Zero Chrome system dialogs afterwards — the extension passively forwards the site's own search responses (read-only by construction;
+- cookies are read by NAME only, values never leave the browser). Until installed, tools return `needs-extension` with the store URL as a clickable link and spend nothing. (A `cdp` fallback via `chrome://inspect` remote debugging still exists for diagnostics: `GOTRY_SESSION_TRANSPORT=cdp` — note Chrome 144+ shows a permission box on every connection.)
 
 ---
 
@@ -141,7 +155,7 @@ Engine verdict:
 ```
 ┌──────────────────────────────────────────────────────────────┐
 │ L1  chat-as-interface; gates are in-message choice cards       │
-│ L2  orchestration  dsh runtime + GoTry plugin (ReAct); 20 tools│
+│ L2  orchestration  dsh runtime + GoTry plugin (ReAct); 21 tools│
 │ L3  domain  unified itinerary model + Z3 feasibility engine    │
 │ L4  data  static packs + hotelbyte-cli bridge + OpenFlights    │
 │ L5  governance  LoopX (objective / gates / evidence / quota)   │
@@ -150,7 +164,7 @@ Engine verdict:
 
 | Layer | Module | Role |
 |---|---|---|
-| L2 | `ts/src/index.ts` (dsh plugin) | 20 tools, time-anchor & memory-brief variables; execute isolation + consent gate + process guards |
+| L2 | `ts/src/index.ts` (dsh plugin) | 21 tools, time-anchor & memory-brief variables; execute isolation + consent gate + per-turn tool budget + process guards |
 | L3 | `ts/src/unified.ts` · `py/gotry_feasibility/` | single solving entry (candidate enumeration + flight-chain Z3) |
 | L4 | `ts/capabilities/effect.ts` · `hbcli.ts` · `skeleton-check.ts` | effect interpreter (backoff retry / circuit breaker / mock interpreter, issue #16) + realtime inventory bridge + OpenFlights skeleton (three-valued semantics) |
 | L5 | loopx governance | objective / gates / evidence / quota |
@@ -161,23 +175,40 @@ Engine verdict:
 
 ## ⚠️ Status & limitations
 
+Evaluation Phase 0 foundation boundary: contracts/registry/validators/unmatched diagnostic fixtures/test-only aggregate admission plus a deterministic PR/nightly/weekly/milestone cadence policy/planner. It returns admission, `pass^k`, budgets, calibration, failure-registry, and cross-benchmark synthesis obligations only; it does not schedule or launch adapters, spend, generate a benchmark score, create an Agent optimization round, or support an uplift claim. No external runner, Python runtime dependency, baseline, or matched production evidence is included.
+
 **Working today** (full-stack regression §1–§34 green; every item has deterministic tests):
 
 - **Z3 solving engine** — feasibility verdicts + door-to-door whole-cost; the historical concurrency race is fixed (§30 regression gate)
 - **Realtime retrieval** — flight/train/hotel (Fliggy official channel), destination/hotel catalogs, weather, live flight observation, route connectivity; realtime prices can overwrite solver prices (`GOTRY_REALTIME_PRICING=1`)
 - **Account session search** — Ctrip flights on your own logged-in Chrome; consent & privacy rules above (see 🔐 **Account session: consent & privacy**)
+- **One-time browser extension setup** — install [GoTry Session Bridge](https://chromewebstore.google.com/detail/gotry-session-bridge/oeajpiccmonococjcegddlooeeohlbgd) from the Chrome Web Store (a browser concern, not a gotry one — one click on "Add to Chrome"). When `gotry_session_search` returns `needs-extension`, the verdict surfaces the store URL as a clickable link in the dsh UI; the background health-watch auto-replays your query once the extension is connected — no manual retry needed.
+- **Extension distribution (issue #21, ADR-21)** — three channels, same extension: the [Chrome Web Store](https://chromewebstore.google.com/detail/gotry-session-bridge/oeajpiccmonococjcegddlooeeohlbgd) (recommended: one-click install + auto-update; live since 2026-09-02) is what dsh users should rely on; the review-free GitHub Releases channel is an opt-in escape hatch; the npm-bundled copy stays the offline-deterministic fallback;
+  - any GitHub-channel failure falls back to the bundled copy. The store re-signs with its own key, so the store build's extension ID differs from the fixed-key unpacked ID — the local bridge trusts both origins (§38 regression).
+- **Session data cross-verification (issues #21 / #67)** — 8 benchmark queries (sf-01..sf-08) verified end-to-end: 7/8 verdict=hit, 6/6 manual-golden soft-score 100%, all hits <15s, zero ReadGuard writes. The comparator is pluggable: `--golden=manual` (default), `--golden=flyai`,
+  - or `--golden=static`. Static mode pins an ODbL OpenFlights route/carrier snapshot and combines it with manual time/price bands;
+  - evidence records requested vs effective source, provenance, estimated fields, and fallback reason. Snapshot/route failure prints a warning to stderr and falls back to manual. Static mode is deterministic benchmark data, **not live schedule, fare, or availability**.
+- **Observed static-source runs (2026-08-30, logged-in Chrome)** — two consecutive runs produced static official 8/8 with zero fallback each time;
+  - Ctrip session hits varied from 3/8 to 5/8, while every scored hit across both runs (3+5 records) passed 13/13 (100%). Non-hits remain explicit `miss` records,
+    - so the ≥90% field score is not presented as 8/8 live availability. The same runs exposed and fixed an online-extension lifecycle bug: idle parked timers/sockets no longer pin the default CLI bridge, while wizardless `keepBridge` behavior remains unchanged (§38: 24/24, §40: 9/9).
 - **Memory & reachability** — motivation profile / wish pool / companions / travel timeline; English output via `GOTRY_LOCALE=en`
+- **Bounded agent tool loops** — a soft convergence context is injected after real dispatch 16; dispatch 18 is the last tool body, already-prepared calls 19+ receive structured `TOOL_BUDGET_EXHAUSTED` failures without entering the body, and inherited native tool schemas are suppressed at `step/end` for a text-only next step. The boundary is exercised through the packaged `dist` entry and a real dsh headless loop against an offline relay; CI packs the current SHA and repeats the E2E from an isolated pnpm consumer install rather than the root development tree. Direct/programmatic calls are outside this per-turn budget.
 
 **Open limitations** (as of 2026-08-29, honest list):
 
 - ⏳ **M3 Exit not closed** — engineering & distribution ready, but real seed-user evidence (50–200 person cohort) not yet accumulated; automated tests prove contracts and formulas, not business pass
 - ⏳ **Ctrip-hotel / Meituan logged-in adapters** — flights done; hotel session surfaces await real login-state backfill (next tick)
 - ⏳ **Interface language** — English currently covers the deterministic solve-output layer only; the dsh host UI and dialogue surface belong to the host / calibration samples
+- ⏳ **External benchmark generalization / Phase 1 bridge** — Round 1's exact DeepSeek treatment was environment-unavailable/schema-invalid (score 0), while GLM timed out at 300 s. Round 2 added the default-off owner-local bridge; its frozen treatment remained diagnostic-only because no structured native bridge call or tagged JSON reached the evaluator. Round 3 added provider-neutral native-call/result/terminal conformance, but its new frozen treatment still stopped after one runner spawn with planner/runner exit 1, zero released terminal bytes, no evaluator entry, and null official scores. Round 4's treatment at SHA `5ebddb2` had primary preflight pass, but planner/runner both exited 1 after 30.968 s, released 0 bytes, the evaluator was not entered, and official scores were null. The product Node gate was v24.20.0 while that treatment used v26.3.0, so the result remains diagnostic-only with no uplift claim. GitHub Node 22/24 §48 separately exposed a source default-off 30 s lifecycle hang. Round 5 is limited to removing the timer/keepalive preload, pinning the root/package DSH closure to alpha.3, making source checkout resolve that locked runtime before the legacy vendored fallback, preserving source normal-mode state under `ts/dsh-runtime/gotry-state/` while benchmark/package runs use the invocation directory for isolation, rejecting a non-alpha.3 benchmark runtime before spawn, enforcing Node 22.15+, and adding a benchmark-only structured diagnostic pipe with allowlisted redacted reason codes while stdout remains fail-closed. Its frozen treatment at code SHA `752e54c` stopped after 140.715 s with `child_nonzero_exit`, zero terminal bytes, and null evaluator/official scores, so it remains diagnostic-only; the lock-consistency successor does not rewrite that UID attribution. Cross-benchmark evidence remains open. See [`docs/benchmark-environment-bridge.md`](docs/benchmark-environment-bridge.md).
+- ⏳ **Round 6 structured terminal diagnostics** — benchmark mode maps only the final structured `turn/end` into closed, redacted model/runtime reason families. A per-session arbiter emits at most once and preserves a specific bridge/conformance reason over a later generic runtime reason; recovered transient retries emit no failure. Free-form messages, raw stderr, paths, prompts, request IDs, and credentials never cross the control pipe. The frozen ChinaTravel treatment at code SHA `c61600b` (`..._00001`, `deepseek-v4-flash`) stopped after 49.546 s with `child_runtime_error`, zero terminal bytes, and null evaluator/official scores; leakage and local credential/endpoint scans were zero. This is diagnostic evidence only and creates no score/uplift claim; later documentation-only successors do not rewrite the treatment attribution.
+- ⏳ **Round 7 minimal benchmark kernel** — at code SHA `edb9392896625adbb48abae4a2ecf968dbfc0349`, benchmark opt-in retains only tool budget, model override, one bridge, and isolation/conformance; product prompt variables, process guards, consent hooks, and ordinary GoTry tools are not installed, while the default path is unchanged. The CLI projects a stable task-agnostic persona and accepts only canonical `insert` and `system-prompt` root items, failing closed on missing, duplicate, quoted, reordered, flow, or noncanonical items. ChinaTravel frozen treatment UID `e20241028160248698752` (`easy`, `deepseek-v4-flash`) passed preflight without fallback, then ended after 80.463 s with runner exit 1 and zero/invalid terminal bytes; the evaluator was not entered, official score is null, and the case is not countable. The allowlisted attribution is `child_bridge_runner_failed`; no uplift or external benchmark closure is claimed. The next problem is a generic bridge-tool schema and recoverable domain-error contract.
 
 <details>
 <summary>📖 Deeper engineering state (ledger contracts / evidence contracts / milestone stance)</summary>
 
-The authoritative state lives in the docs, not this README: transactional state ledger (ADR-15) + dual-form freeze (ADR-16: one ledger semantics for local+web); the M3 real-cohort evidence contract stands (fixtures don't count toward Exit; 50–200 real samples open the gate); the M4 paired-cohort value evidence contract (run-all §34 — synthetic data is never Exit evidence); async work-order terminal contract (`gotry_async_terminal.v1`: 4/4 → succeeded / ledger settled / exit 0). Details: [`docs/roadmap.md`](docs/roadmap.md) / [`docs/architecture.md`](docs/architecture.md) §1 and issues #19–#22.
+The authoritative state lives in the docs, not this README: transactional state ledger (ADR-15) + dual-form freeze (ADR-16: one ledger semantics for local+web); the M3 real-cohort evidence contract stands (fixtures don't count toward Exit; 50–200 real samples open the gate);
+- the M4 paired-cohort value evidence contract (run-all §34 — synthetic data is never Exit evidence);
+- async work-order terminal contract (`gotry_async_terminal.v1`: 4/4 → succeeded / ledger settled / exit 0). Details: [`docs/roadmap.md`](docs/roadmap.md) / [`docs/architecture.md`](docs/architecture.md) §1 and issues #19–#22.
 
 </details>
 
@@ -187,9 +218,16 @@ The authoritative state lives in the docs, not this README: transactional state
 
 ```bash
 ./scripts/run-all-tests.sh
+# Evaluation Phase 0 diagnostic contracts and deterministic cadence planner
+# (offline; no adapter/scheduler/runner/spend/score/Python)
+cd ts
+npx tsx scripts/evaluation-contract-tests.ts
+npx tsx scripts/evaluation-cadence-tests.ts
 ```
 
-One-shot full-stack green (pure TS, no Python needed): golden engines · dialogue replay · cross-process async work-orders · plugin smoke · hbcli · process guards · weather · flights · Anything · probePoi · agent-reach · dual-path stability · time-awareness eval · memory domain · **Z3 race (§30) · realtime pricing (§31) · i18n catalog (§32) · M3 cohort evidence contract (§33) · M4 value evidence contract (§34) · M3 nightly evidence producer contract (§35)**.
+One-shot full-stack green (pure TS, no Python needed): golden engines · dialogue replay · cross-process async work-orders · plugin smoke · hbcli · process guards · weather · flights · Anything · probePoi · agent-reach · dual-path stability · time-awareness eval · memory domain · **Z3 race (§30)
+- · realtime pricing (§31) · i18n catalog (§32) · M3 cohort evidence contract (§33) · M4 value evidence contract (§34) · M3 nightly evidence producer contract (§35) · session transport extension bridge (§38) · extension onboarding pre-check (§40) · bookable-fact gate (§39)
+- · extension distribution channel (§43) · sf-live static-golden offline contracts (§44) · evaluation foundation and cadence policy (§45–§46) · agent tool-budget Cordis integration + dsh headless E2E (§47)**. The live runner remains `cd ts && npx tsx scripts/sf-live-benchmark.ts --golden=static` and requires the user's connected Chrome session.
 
 ---
 
@@ -207,6 +245,6 @@ Standard open-source flow: branch off latest `main` (`feat/ · fix/ · docs/ ·
 
 ---
 
-**Built with**: DeepSeek Harness 0.1.2-alpha.1 (vendored) · Cordis · Z3 (WASM) · loopx (pipx) · hotelbyte-cli · Agent-Reach v1.5.0 (`.venv/`) · OpenFlights · TypeScript
+**Built with**: DeepSeek Harness 0.1.2-alpha.3 (root-pinned) · Cordis · Z3 (WASM) · loopx (pipx) · hotelbyte-cli · Agent-Reach v1.5.0 (`.venv/`) · OpenFlights · TypeScript
 
-**Last verified against `v0.0.1-rc.15` (2026-08-29)** — full-stack regression green §1-§36 (release flow: `scripts/publish-npm.sh`).
+**Version baseline: `v0.0.1-rc.16` (2026-08-30).** The current checkout's authoritative verification gates are enumerated by `scripts/run-all-tests.sh` (release flow: `scripts/publish-npm.sh`).

package/README.zh-CN.md

@@ -18,16 +18,21 @@
 ```bash
 npx @danceiny/gotry web
 # 首跑会提示建 .env:LLM_API_KEY=<DeepSeek key 或 OpenAI 兼容 key>
+#   不走 DeepSeek 官方?再配一行 LLM_BASE_URL=<你的端点,一般以 /v1 结尾>
+#   要指定模型(中转而常需要)?配 LLM_MODEL=<模型名>——同时驱动 dsh 会话面
+#   (web/headless)与仓内脚本,并压过你在 dsh web 设置里选过的模型。
 # → 浏览器打开 http://127.0.0.1:3080,像聊天一样说「我想去大理三天」
 ```
 
+> **换模型 / 换中转?** 价表(`ts/data/llm-price-table.json`,schema `gotry_llm_price_table_v2`)是 `gotry_m3_nightly_run_v1.cost_usd` 唯一事实源(ADR-11)。新增模型或换中转=改本文件走 PR(peak 保守上界只高不低),未知模型 → **fail-closed 不猜价**。漂移监测:`npx tsx ts/scripts/price-drift-watch.ts`(默认离线对照 baseline 比对输出 PR-就绪 Markdown diff;`--fetch` 拉取官方页 + 首次写 fixture)。**永不自动 apply 价格**。
+
 | 你想要的 | 命令 |
 |---|---|
 | 🖥️ 对话规划(推荐) | `npx @danceiny/gotry web` → :3080 对话界面 |
 | 🤖 脚本/一次性问答 | `npx @danceiny/gotry "我想从深圳休整两天,预算 3000"` |
 | 🛠️ 开发者:仓内运行 | 见下方[源码安装](#-快速开始) |
 
-- 前置:Node 22+;一个 LLM API key。零成本启动——dsh 运行时以 cordis patch 自动挂载,无额外配置。
+- 前置:Node 22.15+;一个 LLM API key。任何 OpenAI 兼容端点(MiniMax/中转/自建网关)也可——在 `.env` 另配 `LLM_BASE_URL`(一般以 `/v1` 结尾,如 `https://api.minimax.io/v1`),请求即走它而非 DeepSeek 默认端点。要锁定模型就配 `LLM_MODEL`(如 `MiniMax-M2`):对 dsh 会话面与仓内脚本同时生效,并压过 dsh web 设置里持久化的模型选择;不配则用 dsh 内置默认(`deepseek-v4-flash`)或你在 web 界面的选择。零成本启动——dsh 运行时以 cordis patch 自动挂载,无额外配置。
 - 例外提示:`:3080` 端口被占时先 `kill <PID>`;首启 6–15 秒属正常冷启动;异常退出会留证据到 `gotry-state/incidents.jsonl`(不静默)。
 
 <details>
@@ -35,9 +40,10 @@ npx @danceiny/gotry web
 
 ```bash
 git clone https://github.com/Danceiny/gotry && cd gotry
-cd ts/dsh-runtime && pnpm install && cd ../..    # ① vendored dsh 0.1.2-alpha.1(一次性)
-cp .env.example .env                              # ② 填 LLM_API_KEY
-./gotry web                                       # ③ 仓内入口,同 npm 形态
+npm ci && npm --prefix ts ci                      # ① 安装 root/TS 锁定依赖闭包
+node scripts/build-dist.mjs                       # ② 构建源码检出的 JS runtime
+cp .env.example .env                              # ③ 填 LLM_API_KEY(非 DeepSeek 官方另配 LLM_BASE_URL;锁定模型另配 LLM_MODEL)
+./gotry web                                       # ④ 仓内入口,同 npm 形态
 ```
 
 | 入口 | 命令 | 什么时候用 |
@@ -46,7 +52,7 @@ cp .env.example .env                              # ② 填 LLM_API_KEY
 | headless 一问一答 | `./gotry "一句任务"` | 脚本 / CI / 定向调试 → stdout |
 | help | `./gotry help` | 三行帮助 |
 
-为什么源码模式要先装 runtime:dsh 以 vendored tarball 进 git(`ts/dsh-runtime/`),npm 一键分发不打包它;源码检出读 `.ts` 源码,装完一次重复可用。
+为什么源码模式先装 root/TS 闭包并构建 dist:源码入口与 npm 包都解析同一组 216 个精确直接依赖的 DSH `0.1.2-alpha.3` closure；manifest、package-lock 与 root pnpm importer 必须暴露同一 216 项名称集合，publish preverify 会拒绝漏钉、混版和 range；源码普通运行的 dsh cwd 仍是 `ts/dsh-runtime/`，因此状态继续落 `ts/dsh-runtime/gotry-state/`；benchmark opt-in 与 npm 包运行使用调用目录隔离。`ts/dsh-runtime/` 的旧 vendored runtime 只保留为非 benchmark legacy 解析兼容，不承诺可运行，也不再是推荐安装路径。
 
 </details>
 
@@ -98,7 +104,7 @@ GoTry: 收到。先把约束记下来——
 
 ---
 
-## 🧰 20 个工具 — Tools
+## 🧰 21 个工具 — Tools
 
 | 组 | 工具 | 干什么 |
 |---|---|---|
@@ -115,6 +121,7 @@ GoTry: 收到。先把约束记下来——
 | | `gotry_wish_pool_add` / `gotry_wish_pool_list` | 「下一次出发」愿望池 + 0..1 条件召回 |
 | | `gotry_companion_save` · `gotry_trip_log` | 同行人档案 / 旅行时间线 |
 | **产物** | `gotry_artifacts_list` / `gotry_artifacts_read` | 发现与查看已生成的产物(异步交付 + 工作目录 markdown),行号文件视图,只读 |
+| **事实闸** | `gotry_fact_gate` | 行程产物交付前闸:每条可下单 claim(航班号/时刻/机场/价格/政策)必须回溯到 exact-date 工具结果(hit/miss 均落账);无法回溯 ⇒ blocked——不得宣称「已验证方案」 |
 | **通用外部** | `gotry_web_search` · `gotry_video_subtitle` · `gotry_github_search` · `gotry_agent_reach` | 网页/字幕/GitHub/全渠道外部信息(经 Agent-Reach) |
 
 ---
@@ -128,7 +135,7 @@ GoTry: 收到。先把约束记下来——
 3. **物理只读** —— ReadGuard 在网络层中止一切写请求(下单/支付在传输层不可达),agent 永不接触凭证与验证码;遇到验证码立即停,交还给你。
 4. **绝不劫持你的浏览器** —— 检索/登录只开自己的独立标签页,登录页置前台、留在你那;例行动测试永不自动开浏览器窗。
 
-> 前置(一次性):日常 Chrome 打开 `chrome://inspect/#remote-debugging` 打开开关(Chrome 144+),并保持 Chrome 在线。未开启时工具返回 `needs-attach` 并给出指引,不消耗执行配额。
+> 前置(一次性):去 [Chrome 应用商店](https://chromewebstore.google.com/detail/gotry-session-bridge/oeajpiccmonococjcegddlooeeohlbgd) 点「添加至 Chrome」——这是浏览器的事,不由 gotry 兜(`gotry setup wizard` 不复存在);想用免审核/更新更快的 GitHub 通道见 [docs/extension-webstore-submission.md](docs/extension-webstore-submission.md)。装完**零系统弹窗**——扩展只被动转发站点自己发出的检索响应(构造上只读;cookie 只读名字,值永不离开浏览器)。未安装时工具返回 `needs-extension` 并把商店 URL 作为 dsh UI 的可点链接给出,不消耗执行配额。(诊断后备:cdp 车道经 `chrome://inspect` 远程调试,`GOTRY_SESSION_TRANSPORT=cdp` 显式开启——注意 Chrome 144+ 每次连接都会弹权限框。)
 
 ---
 
@@ -137,7 +144,7 @@ GoTry: 收到。先把约束记下来——
 ```
 ┌──────────────────────────────────────────────────────────────┐
 │ L1  对话即界面  chat-as-UI; gates 是消息内选择题                 │
-│ L2  编排  dsh 运行时 + GoTry 插件(ReAct);20 个工具            │
+│ L2  编排  dsh 运行时 + GoTry 插件(ReAct);21 个工具            │
 │ L3  领域  统一行程模型 + Z3 可行性引擎(枚举/Z3 双形态)        │
 │ L4  数据  静态数据包 + hotelbyte-cli 实时桥 + OpenFlights 骨架 │
 │ L5  治理  LoopX(objective / gates / evidence / quota)         │
@@ -146,7 +153,7 @@ GoTry: 收到。先把约束记下来——
 
 | 层 | 模块 | 角色 |
 |---|---|---|
-| L2 | `ts/src/index.ts`(dsh 插件) | 注册 20 工具,挂时间锚点/记忆 brief 变量;execute 异常隔离 + 授权闸 + 进程护栏 |
+| L2 | `ts/src/index.ts`(dsh 插件) | 注册 21 工具,挂时间锚点/记忆 brief 变量;execute 异常隔离 + 授权闸 + 进程护栏 |
 | L3 | `ts/src/unified.ts` · `py/gotry_feasibility/` | 唯一求解入口(候选枚举 + 航班链 Z3) |
 | L4 | `ts/capabilities/effect.ts` · `hbcli.ts` · `skeleton-check.ts` | 效应解译层(退避重试/断路器/mock 解译器,issue #16)+ 实时库存桥 + OpenFlights 骨架(三值语义) |
 | L5 | loopx 治理面 | objective / gates / evidence / quota |
@@ -162,6 +169,10 @@ GoTry: 收到。先把约束记下来——
 - **Z3 求解引擎** —— 可行性判定 + 门到门全成本;历史并发竞态已根治(§30 并发回归闸)
 - **实时检索**:机票/火车/酒店(飞猪官方通道)、目的地/酒店目录、天气、航班观测、通航性校验;可选让实时票价覆写进求解(`GOTRY_REALTIME_PRICING=1`)
 - **账号会话检索**:你本人登录态查携程机票,授权与隐私规则见上方 **🔐 账号会话:授权与隐私** 小节
+- **一次性浏览器扩展安装**:去 [Chrome 应用商店](https://chromewebstore.google.com/detail/gotry-session-bridge/oeajpiccmonococjcegddlooeeohlbgd) 点「添加至 Chrome」(浏览器的事,gotry 不介入);扩展未装时 `gotry_session_search` 返回的 `needs-extension` 由 dsh UI 把商店 URL 渲成可点链接;后台 health-watch 探活,扩展一就位自动重放你的检索,**无需你手动重跑命令**
+- **扩展分发三通道(issue #21,ADR-21)**:同一扩展三个通道——[Chrome 应用商店](https://chromewebstore.google.com/detail/gotry-session-bridge/oeajpiccmonococjcegddlooeeohlbgd)(dsh 用户就走这条:一键装 + 自动更新,2026-09-02 上架);GitHub Releases 通道(免审核、版本更新更快,见 docs/extension-webstore-submission.md);包内副本保离线兜底。商店用自己签名 key 重签,商店版扩展 ID 与 unpacked 固定 ID 不同,本机桥双 Origin 白名单同信(§38 回归)。
+- **会话数据交叉验证(issue #21 / #67)**:8 条 sf-01..sf-08 benchmark query 端到端验证 —— 7/8 verdict=hit / 6/6 manual-golden 软命中 100% / hit 全部 <15s / ReadGuard 零写。对照源可插拔:`--golden=manual`(默认)、`--golden=flyai`、`--golden=static`。static 模式以 OpenFlights ODbL 固定修订提供航线/承运人,叠加手工时刻与价格带；evidence 分开记录请求源/实际源、provenance、estimated fields 与 fallback reason。快照/路由异常会向 stderr 告警后回退 manual。static 是确定性 benchmark 数据,**不是实时班期、票价或库存**。
+- **static 源登录态观测(2026-08-30)**:连续两轮 static official 均 8/8 hit、零 fallback;携程 session hit 从 3/8 波动到 5/8,两轮所有可评分 hit(3+5 条)全部 13/13(100%)。非 hit 均是显式 `miss`,因此 ≥90% 字段分不表述为 8/8 实时可售。同轮修复扩展在线时空闲 parked timer/socket 钉住默认 CLI 桥的生命周期缺口,wizardless `keepBridge` 语义不变(§38:24/24,§40:9/9)。
 - **记忆与触达**:动机画像 / 愿望池 / 同行人 / 旅行时间线;英文输出一键切换(`GOTRY_LOCALE=en`)
 
 **已知限制**(截至 2026-08-29,诚实清单):
@@ -169,6 +180,9 @@ GoTry: 收到。先把约束记下来——
 - ⏳ **M3 Exit 未关闭** —— 工程与分发面就绪,但"真实种子用户"证据(50–200 人 cohort)尚未积累;自动化测试证明的是合同与公式,不是 business pass
 - ⏳ **携程酒店 / 美团的登录态会话适配** —— 机票已通,酒店面等登录态实测回填(下一个 tick)
 - ⏳ **界面语言** —— 英文界面仅覆盖求解确定性输出层;dsh 宿主界面与对话面属宿主/校准件
+- ⏳ **外部 benchmark 泛化 / Phase 1 bridge** —— Round 1 的 exact DeepSeek treatment 为 environment unavailable/schema-invalid（score 0），GLM 为 300 秒超时。Round 2 建立 default-off owner-local bridge；其冻结 treatment 因没有 structured native bridge call 与 tagged JSON，只到 diagnostic-only。Round 3 增加 provider-neutral 的 native call/result/terminal conformance，但新冻结 treatment 仍在 runner 仅派发一次后以 planner/runner exit 1 停止，parent 释放 0 字节终态，evaluator 未进入，official scores 全为 null。Round 4 treatment（SHA `5ebddb2`）primary preflight=pass，但 planner/runner 均在 30.968s 以 exit 1 结束，释放 0 字节，evaluator 未进入，official scores 全为 null。产品 gate 使用 Node v24.20.0，而该 treatment 使用 v26.3.0，因此该结果仅 diagnostic-only，不产生 uplift。GitHub Node 22/24 §48 另暴露 source default-off 30s lifecycle hang。Round 5 仅限移除 timer/keepalive preload、将 root/package DSH closure 锁定为 alpha.3、让 source checkout 优先解析该锁定 runtime、保留源码普通运行的 `ts/dsh-runtime/gotry-state/` 状态连续性、让 benchmark/package 用调用目录隔离、在 spawn 前拒绝非 alpha.3 benchmark runtime、强制 Node 22.15+，并增加 benchmark-only 结构化诊断 pipe（allowlisted redacted reason codes），stdout 继续 fail-closed；其 frozen treatment（代码 SHA `752e54c`）在 140.715 秒后以 `child_nonzero_exit`、0 terminal bytes、evaluator/official scores null 停止，仅 diagnostic-only，后续纯 lock-consistency 提交不改写该 UID 归属。跨 benchmark 证据仍待运行，不声称 external benchmark closure。详见 [`docs/benchmark-environment-bridge.md`](docs/benchmark-environment-bridge.md)。
+- ⏳ **Round 6 结构化终态诊断** —— benchmark 模式只把最终结构化 `turn/end` 投影为闭合、脱敏的模型/运行时原因族；per-session 仲裁最多写一次，且 bridge/conformance 专项原因优先于后续 generic runtime。恢复成功的 transient retry 不写失败；message、raw stderr、路径、prompt、request id 与凭证不跨控制通道。代码 SHA `c61600b` 的 ChinaTravel 冻结 treatment（`..._00001`，`deepseek-v4-flash`）在 49.546 秒后以 `child_runtime_error`、0 terminal bytes、evaluator/official scores null 停止；leakage 与本地凭证/端点扫描均为 0。该证据只改善归因，不产生 score/uplift 结论；后续纯文档 successor 不改写 treatment 归属。
+- ⏳ **Round 7 minimal kernel** —— 代码 SHA `edb9392896625adbb48abae4a2ecf968dbfc0349` 的 benchmark opt-in 仅保留 tool budget、model override、唯一 bridge、isolation/conformance；不装产品 prompt variables、process guards、consent hooks 与普通 GoTry tools，默认路径不变。CLI 将系统人格投影为稳定、task-agnostic persona，root patch 只接受 canonical `insert` 与 `system-prompt` 各一，缺失、重复、quoted、reordered、flow 或其他非 canonical root item 均 fail-closed。ChinaTravel 冻结 treatment UID `e20241028160248698752`（`easy`，`deepseek-v4-flash`）preflight pass 且未回退，80.463 秒后 runner exit 1、terminal 0 bytes/invalid，evaluator 未进入、official null、不可计分；白名单归因为 `child_bridge_runner_failed`，不产生 uplift 或 external benchmark closure。下一轮聚焦 generic bridge tool schema 与可恢复 domain-error contract。
 
 <details>
 <summary>📖 更深的工程状态(账本合同 / 证据合同 / 里程碑口径)</summary>
@@ -200,7 +214,7 @@ GoTry: 收到。先把约束记下来——
 ./scripts/run-all-tests.sh
 ```
 
-全栈一次性绿(纯 TS,无 Python 依赖):engine/journey/unified 金标准 · 对话重放 · 异步工单跨进程 · 插件 smoke · hbcli · 进程护栏(含工具异常隔离)· 天气 · 航班 · Anything · probePoi · agent-reach(web/deep/wrapper)· 双路径稳定性 · 时间感评测(锚点卡/槽位过期校验/评分器/mock 回放;真模型巡检 `time-eval-tests.ts --real`)· 记忆域(动机合并守门/效用 sidecar/只读指标投影)· **Z3 并发竞态(§30)· 实时票价桥(§31)· i18n 目录(§32)· M3 cohort 证据合同(§33)· M4 价值证据合同(§34)**。
+全栈一次性绿(纯 TS,无 Python 依赖):engine/journey/unified 金标准 · 对话重放 · 异步工单跨进程 · 插件 smoke · hbcli · 进程护栏(含工具异常隔离)· 天气 · 航班 · Anything · probePoi · agent-reach(web/deep/wrapper)· 双路径稳定性 · 时间感评测(锚点卡/槽位过期校验/评分器/mock 回放;真模型巡检 `time-eval-tests.ts --real`)· 记忆域(动机合并守门/效用 sidecar/只读指标投影)· **Z3 并发竞态(§30)· 实时票价桥(§31)· i18n 目录(§32)· M3 cohort 证据合同(§33)· M4 价值证据合同(§34)· 会话传输扩展桥(§38)· 扩展 onboarding 静态预检(§40)· 可下单事实闸(§39)· 扩展分发通道(§43)· sf-live static-golden 离线合同(§44)**。真实 runner 为 `cd ts && npx tsx scripts/sf-live-benchmark.ts --golden=static`,仍需用户 Chrome 会话扩展在线。
 
 ---
 
@@ -230,6 +244,6 @@ GoTry: 收到。先把约束记下来——
 
 ---
 
-**Built with**: DeepSeek Harness 0.1.2-alpha.1 (vendored) · Cordis · Z3 (WASM) · loopx (pipx) · hotelbyte-cli · Agent-Reach v1.5.0 (`.venv/`) · OpenFlights · TypeScript
+**Built with**: DeepSeek Harness 0.1.2-alpha.3 (root-pinned) · Cordis · Z3 (WASM) · loopx (pipx) · hotelbyte-cli · Agent-Reach v1.5.0 (`.venv/`) · OpenFlights · TypeScript
 
-**Last verified against `v0.0.1-rc.15`(2026-08-29)** — 全栈回归全绿 §1-§36(发布流程见 `scripts/publish-npm.sh`)。
+**版本基线:`v0.0.1-rc.16`(2026-08-30)。** 当前 checkout 的权威验证闸以 `scripts/run-all-tests.sh` 实际分节为准(发布流程见 `scripts/publish-npm.sh`)。

package/bin/gotry-booking-copilot.js

@@ -0,0 +1,53 @@
+#!/usr/bin/env node
+/** Standalone, BFF-only Booking Copilot HTTP/SSE service. */
+
+import { existsSync, readFileSync } from 'node:fs'
+import { dirname, join } from 'node:path'
+import { fileURLToPath, pathToFileURL } from 'node:url'
+
+const here = dirname(fileURLToPath(import.meta.url))
+const root = join(here, '..')
+const startupPath = join(root, 'dist/src/booking-surface/startup.js')
+
+function safeError(error) {
+  let message = error instanceof Error ? error.message : String(error)
+  for (const secret of [
+    process.env.GOTRY_BOOKING_COPILOT_API_KEY,
+    process.env.DEEPSEEK_API_KEY,
+    process.env.LLM_API_KEY,
+  ]) {
+    if (secret) message = message.replaceAll(secret, '[REDACTED]')
+  }
+  return message.slice(0, 512)
+}
+
+async function main() {
+  if (!existsSync(startupPath)) throw new Error('booking_copilot_compiled_runtime_missing')
+  const artifactPath = join(root, 'ARTIFACT_ID')
+  if (existsSync(artifactPath)) {
+    const artifactId = readFileSync(artifactPath, 'utf8').trim()
+    if (!/^[0-9a-f]{40}$/.test(artifactId)) throw new Error('booking_copilot_artifact_id_invalid')
+    process.env.GOTRY_BOOKING_COPILOT_ARTIFACT_ID = artifactId
+  }
+  const startup = await import(pathToFileURL(startupPath).href)
+  const config = startup.resolveBookingCopilotStartupConfig(process.env)
+  const handle = await startup.startBookingCopilotFromEnvironment(process.env)
+  process.stderr.write(`[gotry-booking-copilot] listening on http://${config.host}:${handle.port}/a2a/booking-copilot/turn\n`)
+
+  let stopping = false
+  const stop = () => {
+    if (stopping) return
+    stopping = true
+    void handle.close().then(() => process.exit(0), (error) => {
+      process.stderr.write(`[gotry-booking-copilot] shutdown failed: ${safeError(error)}\n`)
+      process.exit(1)
+    })
+  }
+  process.once('SIGINT', stop)
+  process.once('SIGTERM', stop)
+}
+
+main().catch((error) => {
+  process.stderr.write(`[gotry-booking-copilot] startup failed: ${safeError(error)}\n`)
+  process.exitCode = 1
+})