@danceiny/gotry 0.0.1-rc.17 → 0.0.1-rc.18
This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
- package/README.md +158 -147
- package/README.zh-CN.md +173 -154
- package/bin/gotry-bootstrap.js +10 -23
- package/bin/gotry-inner.js +2 -18
- package/dist/scripts/bootstrap-tests.js +29 -43
- package/package.json +1 -1
package/README.md
CHANGED
|
@@ -2,95 +2,74 @@
|
|
|
2
2
|
|
|
3
3
|
# GoTry
|
|
4
4
|
|
|
5
|
-
[](https://github.com/Danceiny/gotry/actions/workflows/ci.yml)
|
|
6
|
-
|
|
7
5
|
> **Body and soul — more travel, less tourism.**
|
|
8
6
|
> *身体和灵魂,更多旅行,更少旅游。*
|
|
9
7
|
|
|
10
|
-
|
|
8
|
+
GoTry is an AI travel agent for **"departure to next departure."** You tell it where you want to go and why; it interviews you about your working hours and existing bookings, then hands you a **formally verified itinerary** — computed by a Z3 solver, not guessed by a model.
|
|
11
9
|
|
|
12
|
-
|
|
13
|
-
|
|
14
|
-
|
|
15
|
-
|
|
16
|
-
|
|
17
|
-
| **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)) |
|
|
10
|
+
[](https://github.com/Danceiny/gotry/actions/workflows/ci.yml)
|
|
11
|
+
[](https://www.npmjs.com/package/@danceiny/gotry)
|
|
12
|
+
[](LICENSE)
|
|
13
|
+
[](https://www.npmjs.com/package/@danceiny/gotry)
|
|
14
|
+
[](docs/architecture.md)
|
|
18
15
|
|
|
19
|
-
|
|
16
|
+
**[What GoTry Does](#what-gotry-does)** · **[How It Works](#how-it-works)** · **[Tools](#tools)** · **[Demo](#demo)** · **[Quick Start](#quick-start)** · **[Consent and Privacy](#consent-and-privacy)** · **[Trustworthy by Construction](#trustworthy-by-construction)** · **[Project Status](#project-status)** · **[Roadmap](#roadmap)** · **[For AI Agents](#for-ai-agents)** · **[Documentation](#documentation)** · **[License](#license)**
|
|
20
17
|
|
|
21
|
-
|
|
18
|
+
> **Tip for newcomers:** one command is enough to feel the difference — `npx @danceiny/gotry web`, open `http://127.0.0.1:3080`, and say *"I want three relaxing days in Dali."* The agent interviews you first; then the solver, not the model, decides what is feasible. Full walkthrough: [`docs/user-guide.md`](docs/user-guide.md).
|
|
22
19
|
|
|
23
|
-
|
|
24
|
-
npx @danceiny/gotry web
|
|
25
|
-
# First run creates .env: LLM_API_KEY=<DeepSeek key, or OpenAI-compatible key>
|
|
26
|
-
# Not using DeepSeek directly? Also set LLM_BASE_URL=<your endpoint, usually ending in /v1>
|
|
27
|
-
# Need a specific model (relays often do)? Set LLM_MODEL=<model name> — it drives the
|
|
28
|
-
# dsh chat face (web/headless) and overrides any model picked in the dsh web settings.
|
|
29
|
-
# → open http://127.0.0.1:3080 and chat: "I want three relaxing days in Dali"
|
|
30
|
-
```
|
|
20
|
+
## What GoTry Does
|
|
31
21
|
|
|
32
|
-
|
|
33
|
-
- . Unknown models **fail-closed** — no guessed prices. Drift monitor: `npx tsx ts/scripts/price-drift-watch.ts` (offline baseline diff;
|
|
34
|
-
- `--fetch` for live official pages). Never auto-applies changes.
|
|
22
|
+
GoTry turns "I want to go somewhere" into "can I — and how, at what true cost?" When the answer is "not this weekend," the destination is caught in a wish pool with its conditions instead of being dropped.
|
|
35
23
|
|
|
36
|
-
|
|
37
|
-
|
|
38
|
-
|
|
39
|
-
| 🤖 Scripted / one-shot answer | `npx @danceiny/gotry "Two recovery days from Shenzhen, budget 3000"` |
|
|
40
|
-
| 🛠️ Developer: run from source | see [source install](#%EF%B8%8F-developer-source-install) below |
|
|
24
|
+
- **For travelers** — a conversational planner that asks the questions that actually matter (working window, booked resources, departure city, budget), then returns a verdict per destination: feasible or not, why, and the **smallest change that makes it feasible**.
|
|
25
|
+
- **For agent builders** — a working example of an agent where the LLM only listens, translates, and explains. Decisions and arithmetic live in a Z3 solver; every deliverable number carries a provenance tag; write operations are gated by design.
|
|
26
|
+
- **Evidence built in** — an estimate never poses as realtime. Tags are attached by the render layer, never by the model, and switch honestly on degradation. Bookable claims that cannot trace to an exact-date tool result are blocked before delivery.
|
|
41
27
|
|
|
42
|
-
|
|
43
|
-
- 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;
|
|
44
|
-
- 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.
|
|
28
|
+
## How It Works
|
|
45
29
|
|
|
46
|
-
|
|
47
|
-
|
|
48
|
-
## ✨ What it does
|
|
49
|
-
|
|
50
|
-
GoTry turns "I want to go somewhere" into "can I, how, and at what true cost":
|
|
30
|
+
One planning pass is a pipeline. The model owns the two language-heavy ends; the solver owns everything numeric:
|
|
51
31
|
|
|
52
32
|
| Stage | Who | Output |
|
|
53
33
|
|---|---|---|
|
|
54
|
-
|
|
|
55
|
-
|
|
|
56
|
-
|
|
|
57
|
-
|
|
|
58
|
-
|
|
|
34
|
+
| Motivation interview | LLM | Mandatory questions: working window / booked resources / departure city |
|
|
35
|
+
| Fact extraction | LLM | Working-hours semantics, leave semantics |
|
|
36
|
+
| Feasibility verdict | **Z3 solver** | Which destinations are feasible / infeasible, why, and the smallest change that makes them feasible |
|
|
37
|
+
| Door-to-door true cost | Solver | Real flight duration (incl. time zones) + early-wake penalty + transfer cost + arrival energy |
|
|
38
|
+
| Evidence chain | Render layer | Every number carries a source tag |
|
|
39
|
+
| Delivery gate | Fact gate | Bookable claims must trace to exact-date tool results, or the artifact is blocked |
|
|
40
|
+
| Memory | Domain layer | Infeasible today → wish pool, with explicit recall conditions |
|
|
59
41
|
|
|
60
|
-
|
|
42
|
+
Vocabulary you will meet in a GoTry answer:
|
|
61
43
|
|
|
62
|
-
|
|
44
|
+
- **Evidence tag** — `[skeleton:openflights]` route existence verified against the public route database; `[realtime:...]` pulled live from an API seconds ago; `[static-pack:estimate]` a researched estimate — not realtime, verify before booking. On degradation the tag switches honestly.
|
|
45
|
+
- **Door-to-door true cost** — the ticket price plus what the trip actually takes from you: real duration across time zones, the early-wake penalty, transfers, and the energy you land with.
|
|
46
|
+
- **Wish pool** — "next departure" storage. An infeasible dream is saved with explicit conditions (e.g. "5+ days, off-season") and recalled when they can be met.
|
|
47
|
+
- **Fact gate** — pre-delivery check on itinerary artifacts: every bookable claim (flight no. / time / airport / price / policy) must trace to an exact-date tool result; unverifiable means blocked — never presented as a verified plan.
|
|
63
48
|
|
|
64
|
-
|
|
49
|
+
Architecture, five layers:
|
|
65
50
|
|
|
66
|
-
### One-liner (npm, recommended)
|
|
67
|
-
|
|
68
|
-
```bash
|
|
69
|
-
npx @danceiny/gotry web
|
|
70
51
|
```
|
|
71
|
-
|
|
72
|
-
|
|
73
|
-
|
|
74
|
-
|
|
75
|
-
|
|
76
|
-
|
|
77
|
-
|
|
78
|
-
npm ci && npm --prefix ts ci # ① install the pinned root/TS closure
|
|
79
|
-
node scripts/build-dist.mjs # ② build the source checkout's JS runtime
|
|
80
|
-
cp .env.example .env # ③ set LLM_API_KEY (+ LLM_BASE_URL if not on DeepSeek; + LLM_MODEL to pin the model)
|
|
81
|
-
./gotry web # ④ in-repo entry, same UX
|
|
52
|
+
┌──────────────────────────────────────────────────────────────┐
|
|
53
|
+
│ L1 chat-as-interface; gates are in-message choice cards │
|
|
54
|
+
│ L2 orchestration dsh runtime + GoTry plugin (ReAct); 21 tools│
|
|
55
|
+
│ L3 domain unified itinerary model + Z3 feasibility engine │
|
|
56
|
+
│ L4 data static packs + hotelbyte-cli bridge + OpenFlights │
|
|
57
|
+
│ L5 governance LoopX (objective / gates / evidence / quota) │
|
|
58
|
+
└──────────────────────────────────────────────────────────────┘
|
|
82
59
|
```
|
|
83
60
|
|
|
84
|
-
|
|
|
61
|
+
| Layer | Module | Role |
|
|
85
62
|
|---|---|---|
|
|
86
|
-
|
|
|
87
|
-
|
|
|
63
|
+
| L2 | `ts/src/index.ts` (dsh plugin) | 21 tools, time-anchor & memory-brief variables; execute isolation + consent gate + per-turn tool budget + process guards |
|
|
64
|
+
| L3 | `ts/src/unified.ts` · `py/gotry_feasibility/` | single solving entry (candidate enumeration + flight-chain Z3) |
|
|
65
|
+
| L4 | `ts/capabilities/effect.ts` · `hbcli.ts` · `skeleton-check.ts` | effect interpreter (backoff retry / circuit breaker / mock interpreter) + realtime inventory bridge + OpenFlights skeleton (three-valued semantics) |
|
|
66
|
+
| L5 | loopx governance | objective / gates / evidence / quota |
|
|
88
67
|
|
|
89
|
-
|
|
68
|
+
> Full ADRs / evolution / debt ledger: [`docs/architecture.md`](docs/architecture.md) (Chinese — English versions planned for v0.1.0).
|
|
90
69
|
|
|
91
|
-
|
|
70
|
+
## Tools
|
|
92
71
|
|
|
93
|
-
|
|
72
|
+
21 tools in six groups:
|
|
94
73
|
|
|
95
74
|
| Group | Tool | What it does |
|
|
96
75
|
|---|---|---|
|
|
@@ -107,26 +86,10 @@ The source entry and npm package both resolve one 216-package DSH `0.1.2-alpha.3
|
|
|
107
86
|
| | `gotry_wish_pool_add` / `gotry_wish_pool_list` | "next departure" wish pool + 0..1 conditional recall |
|
|
108
87
|
| | `gotry_companion_save` · `gotry_trip_log` | companion profile / travel timeline |
|
|
109
88
|
| **Artifacts** | `gotry_artifacts_list` / `gotry_artifacts_read` | Discover & view generated artifacts (async deliverables + working-dir markdown) as a line-numbered file view (read-only) |
|
|
110
|
-
| **Factuality gate** | `gotry_fact_gate` | Pre-delivery gate for itinerary artifacts
|
|
89
|
+
| **Factuality gate** | `gotry_fact_gate` | Pre-delivery gate for itinerary artifacts — see [fact gate](#how-it-works) above |
|
|
111
90
|
| **General external** | `gotry_web_search` · `gotry_video_subtitle` · `gotry_github_search` · `gotry_agent_reach` | web / subtitles / GitHub / all-channel external info (via Agent-Reach) |
|
|
112
91
|
|
|
113
|
-
|
|
114
|
-
|
|
115
|
-
## 🔐 Account session: consent & privacy
|
|
116
|
-
|
|
117
|
-
The account session channel reads realtime hotel/flight data from **the user's own logged-in Chrome**, under four hard rules:
|
|
118
|
-
|
|
119
|
-
1. **Login happens on the external website.** GoTry never offers, fills, or collects any password / SMS code / cookie value. It only answers one boolean question: "does a login-ticket cookie exist" (reads cookie **names** only — zero values touched).
|
|
120
|
-
2. **Consent card, once per session.** The first account-session use pops a runtime approval card; approval holds for the session, a refusal revokes it for the session (no repeat prompting). Master switch `sessionAccess: ask|allow|off` at any time.
|
|
121
|
-
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.
|
|
122
|
-
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.
|
|
123
|
-
|
|
124
|
-
> 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;
|
|
125
|
-
- 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.)
|
|
126
|
-
|
|
127
|
-
---
|
|
128
|
-
|
|
129
|
-
## 🎬 A conversation — Demo
|
|
92
|
+
## Demo
|
|
130
93
|
|
|
131
94
|
```
|
|
132
95
|
> Two or three days staring at Erhai Lake, leaving from Shanghai, budget 3000, annual leave — no work.
|
|
@@ -146,105 +109,153 @@ Engine verdict:
|
|
|
146
109
|
[static-pack:estimate] G7315/G7316 priced on Jul–Aug off-season rates
|
|
147
110
|
```
|
|
148
111
|
|
|
149
|
-
> Tag guide: `[
|
|
112
|
+
> Tag guide: `[skeleton:openflights]` means "this route can be flown" was verified against the public route database; `[realtime:...]` marks data pulled live seconds ago; `[static-pack:estimate]` flags an off-season estimate — **verify before booking**. Tags are attached by the render layer, never by the model.
|
|
150
113
|
|
|
151
|
-
|
|
114
|
+
## Quick Start
|
|
152
115
|
|
|
153
|
-
|
|
116
|
+
### npm (recommended)
|
|
154
117
|
|
|
155
|
-
```
|
|
156
|
-
|
|
157
|
-
|
|
158
|
-
|
|
159
|
-
│ L3 domain unified itinerary model + Z3 feasibility engine │
|
|
160
|
-
│ L4 data static packs + hotelbyte-cli bridge + OpenFlights │
|
|
161
|
-
│ L5 governance LoopX (objective / gates / evidence / quota) │
|
|
162
|
-
└──────────────────────────────────────────────────────────────┘
|
|
118
|
+
```bash
|
|
119
|
+
npx @danceiny/gotry web
|
|
120
|
+
# → open http://127.0.0.1:3080 and chat: "I want three relaxing days in Dali"
|
|
121
|
+
# LLM key & model: handled by the dsh host UI; nothing for gotry to ask on the CLI
|
|
163
122
|
```
|
|
164
123
|
|
|
165
|
-
|
|
|
124
|
+
| Entry | Command | When |
|
|
166
125
|
|---|---|---|
|
|
167
|
-
|
|
|
168
|
-
|
|
|
169
|
-
| 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) |
|
|
170
|
-
| L5 | loopx governance | objective / gates / evidence / quota |
|
|
126
|
+
| Web chat (recommended) | `npx @danceiny/gotry web` | multi-turn planning with visualized reasoning → `:3080` |
|
|
127
|
+
| Headless one-shot | `npx @danceiny/gotry "Two recovery days from Shenzhen, budget 3000"` | scripts / CI / targeted debugging → stdout |
|
|
171
128
|
|
|
172
|
-
|
|
129
|
+
Requires Node ≥ 22.15. LLM credentials are managed by your dsh host UI — gotry itself never asks for or echoes them. OpenAI-compatible endpoints (MiniMax / relays / self-hosted gateways) are handled by the dsh model configuration. First cold start takes 6–15 s; if port `:3080` is taken, free it first; unexpected exits leave evidence in `gotry-state/incidents.jsonl` (nothing silent).
|
|
173
130
|
|
|
174
|
-
|
|
131
|
+
> **Cost accounting** — `ts/data/llm-price-table.json` (schema `gotry_llm_price_table_v2`) is the single source of truth for nightly run cost. Adding a model or switching relays = a PR against this file (peak-conservative upper bounds 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). It never auto-applies changes.
|
|
175
132
|
|
|
176
|
-
|
|
133
|
+
### Developer source install
|
|
134
|
+
|
|
135
|
+
```bash
|
|
136
|
+
git clone https://github.com/Danceiny/gotry && cd gotry
|
|
137
|
+
npm ci && npm --prefix ts ci # pinned root/TS closure
|
|
138
|
+
node scripts/build-dist.mjs # build the JS runtime
|
|
139
|
+
./gotry web # in-repo entry, same UX
|
|
140
|
+
```
|
|
141
|
+
|
|
142
|
+
The source entry and the npm package resolve the same 216-package DeepSeek Harness `0.1.2-alpha.3` closure (exact direct dependencies; publish preverify rejects omissions, mixed versions, and ranges). Source normal runs keep their state under `ts/dsh-runtime/gotry-state/`; benchmark opt-in and npm-package runs use the invocation directory for isolation.
|
|
143
|
+
|
|
144
|
+
## Consent and Privacy
|
|
145
|
+
|
|
146
|
+
The account-session channel reads realtime hotel/flight data from **your own logged-in Chrome**, under four hard rules:
|
|
147
|
+
|
|
148
|
+
1. **Login happens on the external website.** GoTry never offers, fills, or collects any password / SMS code / cookie value. It only answers one boolean question — "does a login-ticket cookie exist" (reads cookie **names** only, zero values touched). Existing logins are auto-detected with zero popups.
|
|
149
|
+
2. **Consent card, once per session.** The first account-session use pops a runtime approval card; approval holds for the session, a refusal revokes it (no repeat prompting). Master switch `sessionAccess: ask|allow|off` at any time.
|
|
150
|
+
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 to you.
|
|
151
|
+
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.
|
|
152
|
+
|
|
153
|
+
One-time prerequisite: the [GoTry Session Bridge](https://chromewebstore.google.com/detail/gotry-session-bridge/oeajpiccmonococjcegddlooeeohlbgd) Chrome extension — handled by the dsh host UI when an account-session tool first needs it (`gotry_session_search` surfaces the install URL as a clickable link in the verdict). The extension itself is one-click on the Chrome Web Store, auto-updates, and the gotry side never asks the user to load unpacked or to run a setup wizard. 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). A background health-watch auto-replays your query once the extension is connected. Until installed, tools return `needs-extension` with the store URL and spend nothing.
|
|
177
154
|
|
|
178
|
-
|
|
155
|
+
## Trustworthy by Construction
|
|
179
156
|
|
|
180
|
-
|
|
157
|
+
1. **The model translates; the solver decides.** The LLM never produces feasibility verdicts or arithmetic — those are computed by Z3 against the extracted facts.
|
|
158
|
+
2. **Every number carries a source tag** — attached by the render layer, never the model. Tags switch honestly on degradation; an estimate never poses as realtime.
|
|
159
|
+
3. **No write path exists.** Booking/payment-class tools must pass WriteGate before any implementation ships; the future booking seam is already pinned by the `booking_saga_fsm.v1` edge table.
|
|
160
|
+
4. **Login never touches credentials.** Login happens on the external website; GoTry reads cookie names only; consent is asked once per session and revocable.
|
|
161
|
+
5. **Retrieval is physically read-only.** A ReadGuard aborts write requests at the network layer; a captcha stops the agent and hands control back to you.
|
|
162
|
+
6. **Unverifiable means blocked.** The fact gate refuses to deliver any itinerary whose bookable claims cannot trace to exact-date tool results — it is never presented as a verified plan.
|
|
163
|
+
7. **Prices fail closed.** Unknown models get no guessed price; the price table changes only by PR; the drift monitor reports, never auto-applies.
|
|
164
|
+
8. **Your data is yours.** Product state lives under `gotry-state/`; automated tests and smoke runs use isolated state roots and never write the founder's real product data.
|
|
181
165
|
|
|
182
|
-
|
|
183
|
-
- **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`)
|
|
184
|
-
- **Account session search** — Ctrip flights on your own logged-in Chrome; consent & privacy rules above (see 🔐 **Account session: consent & privacy**)
|
|
185
|
-
- **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.
|
|
186
|
-
- **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;
|
|
187
|
-
- 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).
|
|
188
|
-
- **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`,
|
|
189
|
-
- or `--golden=static`. Static mode pins an ODbL OpenFlights route/carrier snapshot and combines it with manual time/price bands;
|
|
190
|
-
- 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**.
|
|
191
|
-
- **Observed static-source runs (2026-08-30, logged-in Chrome)** — two consecutive runs produced static official 8/8 with zero fallback each time;
|
|
192
|
-
- 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,
|
|
193
|
-
- 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).
|
|
194
|
-
- **Memory & reachability** — motivation profile / wish pool / companions / travel timeline; English output via `GOTRY_LOCALE=en`
|
|
195
|
-
- **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.
|
|
166
|
+
## Project Status
|
|
196
167
|
|
|
197
|
-
|
|
168
|
+
Current release: **v0.0.1-rc.17** (npm dist-tag `rc`; `latest` stays on rc.16 per the dsh-as-LLM-host install convention). Evaluation is at Phase 0 foundation — deterministic contracts, validators, and a cadence policy; no external benchmark scores, no spend, no uplift claims.
|
|
198
169
|
|
|
199
|
-
|
|
200
|
-
|
|
201
|
-
-
|
|
202
|
-
-
|
|
203
|
-
-
|
|
204
|
-
-
|
|
170
|
+
**Working today** (full-stack regression green; every item has deterministic tests):
|
|
171
|
+
|
|
172
|
+
- **Z3 solving engine** — feasibility verdicts + door-to-door whole-cost; the historical concurrency race is fixed and regression-gated
|
|
173
|
+
- **Realtime retrieval** — flights/trains/hotels (Fliggy official channel), destination/hotel catalogs, weather, live flight observation, route connectivity; realtime prices can overwrite solver prices (`GOTRY_REALTIME_PRICING=1`)
|
|
174
|
+
- **Account-session search** — Ctrip flights on your logged-in Chrome; observed runs scored every landed hit 13/13 with zero write attempts, while non-hits stay explicit `miss` records — no live-availability claim beyond that
|
|
175
|
+
- **Extension install on demand** — `[GoTry Session Bridge](https://chromewebstore.google.com/detail/gotry-session-bridge/oeajpiccmonococjcegddlooeeohlbgd)` is offered as a clickable link in the dsh UI when an account-session tool first needs it (one-click install + auto-update); the gotry side never runs a setup wizard
|
|
176
|
+
- **Memory & reachability** — motivation profile / wish pool / companions / travel timeline; English solve output via `GOTRY_LOCALE=en`
|
|
177
|
+
- **Bounded agent tool loops** — soft convergence after 16 dispatches, structured `TOOL_BUDGET_EXHAUSTED` refusals beyond 18, exercised end-to-end through a packaged consumer install in CI
|
|
178
|
+
|
|
179
|
+
**Open limitations** (honest list):
|
|
180
|
+
|
|
181
|
+
- **M3 Exit not closed** — engineering & distribution are ready, but real seed-user evidence (50–200 person cohort) has not been accumulated; automated tests prove contracts and formulas, not business pass
|
|
182
|
+
- **Hotel session adapters** — Ctrip-hotel / Meituan logged-in surfaces await real login-state backfill; flights are done
|
|
183
|
+
- **Interface language** — English covers the deterministic solve-output layer; the dsh host UI and dialogue surface belong to the host / calibration samples
|
|
184
|
+
- **External benchmark generalization** — every frozen external run to date remains diagnostic-only (no score, no uplift claim); the round-by-round engineering ledger lives in [`docs/benchmark-environment-bridge.md`](docs/benchmark-environment-bridge.md)
|
|
185
|
+
- **Booking** — nothing bookable ships today; M5 opens only through WriteGate and the booking-saga FSM
|
|
205
186
|
|
|
206
187
|
<details>
|
|
207
|
-
<summary
|
|
188
|
+
<summary>Deeper engineering state (ledger contracts / evidence contracts / milestone stance)</summary>
|
|
208
189
|
|
|
209
|
-
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);
|
|
210
|
-
- the M4 paired-cohort value evidence contract (run-all §34 — synthetic data is never Exit evidence);
|
|
211
|
-
- 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.
|
|
190
|
+
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 (synthetic data is never Exit evidence); the 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.
|
|
212
191
|
|
|
213
192
|
</details>
|
|
214
193
|
|
|
215
|
-
|
|
194
|
+
## Roadmap
|
|
195
|
+
|
|
196
|
+
| # | Milestone | Scope | Status |
|
|
197
|
+
|---|---|---|---|
|
|
198
|
+
| M0 | Deterministic pipeline | dual engine implementations + real data packs + reconciliation | ✅ |
|
|
199
|
+
| M1 | Agent form established | LLM in the loop; chat as interface; gates as choice cards | ✅ 2026-08-22 |
|
|
200
|
+
| M2 | Realtime data | hotelbyte bridge + flight sources; evidence chain switches to realtime tags | ✅ 2026-08-22 |
|
|
201
|
+
| M3 | MVP | minimal web face + 50–200 seed users (Erhai / Phuket scenarios) | **← current — evidence open** |
|
|
202
|
+
| M4 | Memory & "next departure" | six-layer memory C-end domain; paired-cohort value evidence | founder-authorized parallel |
|
|
203
|
+
| M5 | Transaction loop | WriteGate in production; booking / payment / refunds | entry-gated |
|
|
204
|
+
| M6 | B2B embedding | principal/sponsor plugin with zero kernel changes | entry-gated |
|
|
216
205
|
|
|
217
|
-
|
|
206
|
+
The single authoritative timeline — entry/exit conditions, deliverables, and gates per milestone — is [`docs/roadmap.md`](docs/roadmap.md).
|
|
207
|
+
|
|
208
|
+
## Verify
|
|
218
209
|
|
|
219
210
|
```bash
|
|
220
|
-
./scripts/run-all-tests.sh
|
|
221
|
-
# Evaluation Phase 0 diagnostic contracts and deterministic cadence planner
|
|
222
|
-
# (offline; no adapter/scheduler/runner/spend/score/Python)
|
|
211
|
+
./scripts/run-all-tests.sh # full-stack suite (pure TS, no Python needed)
|
|
223
212
|
cd ts
|
|
224
|
-
npx tsx scripts/evaluation-contract-tests.ts
|
|
225
|
-
npx tsx scripts/evaluation-cadence-tests.ts
|
|
213
|
+
npx tsx scripts/evaluation-contract-tests.ts # evaluation Phase 0 contracts (offline)
|
|
214
|
+
npx tsx scripts/evaluation-cadence-tests.ts # deterministic cadence policy/planner
|
|
226
215
|
```
|
|
227
216
|
|
|
228
|
-
|
|
229
|
-
- · 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)
|
|
230
|
-
- · 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.
|
|
217
|
+
The suite covers golden engines, dialogue replay, cross-process async work-orders, plugin smoke, realtime bridges, process guards, i18n, memory domain, the Z3 concurrency gate, the fact gate, and a packaged-consumer tool-budget E2E, among others; the authoritative section list is whatever `scripts/run-all-tests.sh` enumerates. The live session benchmark (`npx tsx scripts/sf-live-benchmark.ts --golden=static`) is opt-in, requires your connected Chrome session, and never runs in CI.
|
|
231
218
|
|
|
232
|
-
|
|
219
|
+
## Contributing
|
|
233
220
|
|
|
234
|
-
|
|
221
|
+
Branch off latest `main` (`feat/ · fix/ · docs/ · chore/`), full suite green locally, open a Pull Request — `main` never takes direct pushes. CI (Node 22/24, typecheck + all suites) plus maintainer review, then squash-merge. **Red tests never merge.** Full guide: [CONTRIBUTING.md](CONTRIBUTING.md). Bug reports / feature suggestions: use the issue templates (search existing issues first).
|
|
235
222
|
|
|
236
|
-
|
|
223
|
+
## For AI Agents
|
|
237
224
|
|
|
238
|
-
|
|
225
|
+
If you are an agent working in this repository, [`AGENTS.md`](AGENTS.md) is the binding contract — read it first. In brief:
|
|
239
226
|
|
|
240
|
-
|
|
227
|
+
- **Sweep async work orders on entry**: `ts/gotry-state/async/*.json` without a matching `.deliverable.md` → `cd ts && npx tsx scripts/async-collect.ts <id>`.
|
|
228
|
+
- **Layer discipline**: arithmetic only in the evaluate layer of `model.ts` / `unified.py`; solving only in `unified.ts` / `unified.py`; `engine.*` / `journey.*` are deprecated compatibility layers — new code must not call them. Any side change requires the full-stack regression.
|
|
229
|
+
- **Never write shared state**: `ts/dsh-runtime/gotry-state/` is the founder's real product data; validate write paths with an isolated `stateRoot` only.
|
|
230
|
+
- **State-sync discipline**: any commit changing the system's shape/state/debt must sync the six state faces of `architecture.md` §11 in the same commit; stage only named files — never `git add -A`.
|
|
241
231
|
|
|
242
|
-
|
|
232
|
+
Program-level context: [`docs/gotry-master-outline.md`](docs/gotry-master-outline.md). Technical authority: [`docs/architecture.md`](docs/architecture.md).
|
|
243
233
|
|
|
244
|
-
|
|
234
|
+
## Documentation
|
|
235
|
+
|
|
236
|
+
| Document | Purpose |
|
|
237
|
+
|---|---|
|
|
238
|
+
| [`docs/architecture.md`](docs/architecture.md) | System, ADRs, evolution, debt ledger (Chinese, authoritative) |
|
|
239
|
+
| [`docs/gotry-master-outline.md`](docs/gotry-master-outline.md) | Program master outline & reuse matrix |
|
|
240
|
+
| [`docs/gotry-product-design.md`](docs/gotry-product-design.md) | Product design: main loop, transparency, whole-cost model |
|
|
241
|
+
| [`docs/roadmap.md`](docs/roadmap.md) | M0–M6 timeline & current position |
|
|
242
|
+
| [`docs/user-guide.md`](docs/user-guide.md) | End-user guide |
|
|
243
|
+
| [`docs/data-sources.md`](docs/data-sources.md) | Data sources & evidence-chain policy |
|
|
244
|
+
| [`docs/extension-privacy.md`](docs/extension-privacy.md) | Session Bridge extension privacy |
|
|
245
|
+
| [`docs/benchmark-environment-bridge.md`](docs/benchmark-environment-bridge.md) | External benchmark bridge — engineering ledger |
|
|
246
|
+
| [`docs/evaluation-foundation.md`](docs/evaluation-foundation.md) | Evaluation Phase 0 foundation |
|
|
247
|
+
| [`docs/booking-saga-fsm.md`](docs/booking-saga-fsm.md) | Booking saga FSM (the M5 seam vocabulary) |
|
|
248
|
+
| [`docs/kimi-postmortem.md`](docs/kimi-postmortem.md) | A real AI-travel-planning failure postmortem (cautionary tale) |
|
|
249
|
+
| [`docs/release-notes.md`](docs/release-notes.md) | Release decisions per version (the "why") |
|
|
250
|
+
| [`CHANGELOG.md`](CHANGELOG.md) | Machine-derived changelog (Keep a Changelog + Conventional Commits) |
|
|
251
|
+
| [`docs/tokens.md`](docs/tokens.md) | npm 2FA / release mechanics |
|
|
252
|
+
|
|
253
|
+
## License
|
|
254
|
+
|
|
255
|
+
**MIT** — same as upstream dsh. See [LICENSE](LICENSE).
|
|
245
256
|
|
|
246
257
|
---
|
|
247
258
|
|
|
248
|
-
**Built with**: DeepSeek Harness 0.1.2-alpha.3 (root-pinned) · Cordis · Z3 (WASM) · loopx (pipx) · hotelbyte-cli · Agent-Reach v1.5.0
|
|
259
|
+
**Built with**: DeepSeek Harness 0.1.2-alpha.3 (root-pinned) · Cordis · Z3 (WASM) · loopx (pipx) · hotelbyte-cli · Agent-Reach v1.5.0 · OpenFlights · TypeScript
|
|
249
260
|
|
|
250
|
-
**Version baseline: `v0.0.1-rc.16` (
|
|
261
|
+
**Version baseline: `v0.0.1-rc.16` (npm `latest`).** The authoritative verification gates for the current checkout are enumerated by `scripts/run-all-tests.sh`; release flow: `scripts/publish-npm.sh`.
|
package/README.zh-CN.md
CHANGED
|
@@ -1,90 +1,101 @@
|
|
|
1
|
-
# GoTry
|
|
2
|
-
|
|
3
1
|
[English](README.md) | [简体中文](README.zh-CN.md)
|
|
4
2
|
|
|
5
|
-
|
|
3
|
+
# GoTry
|
|
6
4
|
|
|
7
5
|
> **身体和灵魂,更多旅行,更少旅游。**
|
|
8
6
|
> *Body and soul — more travel, less tourism.*
|
|
9
7
|
|
|
10
|
-
**GoTry 是「从出发到下一次出发」的 AI 旅行 Agent
|
|
8
|
+
**GoTry 是「从出发到下一次出发」的 AI 旅行 Agent**:你用一句话说想去哪、为什么想出发;它先问清楚你的工作窗口和已订资源,再用 **Z3 数学求解器**给你一份经过形式化验证的行程方案——是算出来的,不是模型猜的。
|
|
11
9
|
|
|
12
|
-
|
|
10
|
+
[](https://github.com/Danceiny/gotry/actions/workflows/ci.yml)
|
|
11
|
+
[](https://www.npmjs.com/package/@danceiny/gotry)
|
|
12
|
+
[](LICENSE)
|
|
13
|
+
[](https://www.npmjs.com/package/@danceiny/gotry)
|
|
14
|
+
[](docs/architecture.md)
|
|
13
15
|
|
|
14
|
-
|
|
16
|
+
**[它做什么](#它做什么)** · **[工作原理](#工作原理)** · **[工具](#工具)** · **[一段对话](#一段对话)** · **[快速开始](#快速开始)** · **[账号会话:授权与隐私](#账号会话授权与隐私)** · **[构造上可信](#构造上可信)** · **[状态与限制](#状态与限制)** · **[路线图](#路线图)** · **[给 AI Agent](#给-ai-agent)** · **[文档](#文档)** · **[License](#license)**
|
|
15
17
|
|
|
16
|
-
|
|
18
|
+
> **新读者提示**:一条命令就能感受差别——`npx @danceiny/gotry web`,打开 `http://127.0.0.1:3080`,说一句「我想去大理躺三天」。Agent 会先访谈你;然后由求解器(而非模型)判定什么可行。完整走查见 [`docs/user-guide.md`](docs/user-guide.md)。
|
|
17
19
|
|
|
18
|
-
|
|
19
|
-
npx @danceiny/gotry web
|
|
20
|
-
# 首跑会提示建 .env:LLM_API_KEY=<DeepSeek key 或 OpenAI 兼容 key>
|
|
21
|
-
# 不走 DeepSeek 官方?再配一行 LLM_BASE_URL=<你的端点,一般以 /v1 结尾>
|
|
22
|
-
# 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
|
|
23
|
-
# (web/headless)与仓内脚本,并压过你在 dsh web 设置里选过的模型。
|
|
24
|
-
# → 浏览器打开 http://127.0.0.1:3080,像聊天一样说「我想去大理三天」
|
|
25
|
-
```
|
|
20
|
+
## 它做什么
|
|
26
21
|
|
|
27
|
-
|
|
22
|
+
GoTry 把「想去哪」变成「能不能——怎么去、真实代价是多少」。当答案是「这周末不行」时,目的地连同成行条件进愿望池接住,而不是被丢掉。
|
|
28
23
|
|
|
29
|
-
|
|
30
|
-
|
|
31
|
-
|
|
32
|
-
| 🤖 脚本/一次性问答 | `npx @danceiny/gotry "我想从深圳休整两天,预算 3000"` |
|
|
33
|
-
| 🛠️ 开发者:仓内运行 | 见下方[源码安装](#-快速开始) |
|
|
24
|
+
- **给旅行者** —— 一个先问对人问题(工作窗口/已订资源/出发城市/预算)的对话规划器,然后逐目的地给判决:可行/不可行、为什么、以及**让它可行的最小改动**。
|
|
25
|
+
- **给 Agent 工程师** —— 一个「LLM 只负责听懂、翻译、解释」的落地样本:判定与算术在 Z3 求解器里,交付物里每个数字都带来源标签,写操作从设计上就是被闸住的。
|
|
26
|
+
- **证据内建** —— 估算绝不冒充实时。标签由渲染层附加、模型无权染指;降级时标签如实更换。回溯不到 exact-date 工具结果的可下单 claim,交付前就被拦下。
|
|
34
27
|
|
|
35
|
-
|
|
36
|
-
- 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)。
|
|
28
|
+
## 工作原理
|
|
37
29
|
|
|
38
|
-
|
|
39
|
-
<summary>🛠️ 开发者:源码安装(仓内运行)</summary>
|
|
30
|
+
一次规划是一条流水线:模型只占语言密集的两端,数值全部归求解器:
|
|
40
31
|
|
|
41
|
-
|
|
42
|
-
git clone https://github.com/Danceiny/gotry && cd gotry
|
|
43
|
-
npm ci && npm --prefix ts ci # ① 安装 root/TS 锁定依赖闭包
|
|
44
|
-
node scripts/build-dist.mjs # ② 构建源码检出的 JS runtime
|
|
45
|
-
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)
|
|
46
|
-
./gotry web # ④ 仓内入口,同 npm 形态
|
|
47
|
-
```
|
|
48
|
-
|
|
49
|
-
| 入口 | 命令 | 什么时候用 |
|
|
32
|
+
| 阶段 | 由谁做 | 给你什么 |
|
|
50
33
|
|---|---|---|
|
|
51
|
-
|
|
|
52
|
-
|
|
|
53
|
-
|
|
|
54
|
-
|
|
55
|
-
|
|
34
|
+
| 动机访谈 | LLM | 必问项:工作窗口 / 已订资源 / 出发城市 |
|
|
35
|
+
| 事实抽取 | LLM | 工作窗口生效 + 休假语义识别 |
|
|
36
|
+
| 可行性判决 | **Z3 求解器** | 哪些候选可行/不可行、为什么、**最小改动让它可行** |
|
|
37
|
+
| 门到门全成本 | 求解器 | 真实飞行时长(含时差)+ 早起惩罚 + 接驳代价 + 到达精力 % |
|
|
38
|
+
| 证据链 | 渲染层 | 每个数字都带来源标签 |
|
|
39
|
+
| 交付闸 | 事实闸 | 可下单 claim 必须回溯到 exact-date 工具结果,否则产物 blocked |
|
|
40
|
+
| 记忆 | 领域层 | 当下不可行 → 愿望池,附显式召回条件 |
|
|
56
41
|
|
|
57
|
-
|
|
42
|
+
在 GoTry 的回答里会遇到的词汇:
|
|
58
43
|
|
|
59
|
-
|
|
44
|
+
- **证据标签** —— `[骨架:openflights]` 航线存在性经公开航线库校验;`[实时API:...]` 刚从实时接口拉回数秒;`[静态包:估算]` 调研估算——非实时,下单前请核实。降级时标签如实更换。
|
|
45
|
+
- **门到门全成本** —— 票价之外,这段旅程真正从你身上拿走的东西:跨时区的真实时长、早起惩罚、接驳、落地时的精力余额。
|
|
46
|
+
- **愿望池** —— 「下一次出发」的存储。装不下的憧憬带显式条件(如「5 天+、淡季」)入池,条件满足时被召回。
|
|
47
|
+
- **事实闸** —— 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;回溯不到即 blocked——绝不宣称「已验证方案」。
|
|
60
48
|
|
|
61
|
-
|
|
49
|
+
架构五层:
|
|
62
50
|
|
|
63
|
-
|
|
51
|
+
```
|
|
52
|
+
┌──────────────────────────────────────────────────────────────┐
|
|
53
|
+
│ L1 对话即界面 chat-as-UI; gates 是消息内选择题 │
|
|
54
|
+
│ L2 编排 dsh 运行时 + GoTry 插件(ReAct);21 个工具 │
|
|
55
|
+
│ L3 领域 统一行程模型 + Z3 可行性引擎 │
|
|
56
|
+
│ L4 数据 静态数据包 + hotelbyte-cli 实时桥 + OpenFlights 骨架 │
|
|
57
|
+
│ L5 治理 LoopX(objective / gates / evidence / quota) │
|
|
58
|
+
└──────────────────────────────────────────────────────────────┘
|
|
59
|
+
```
|
|
64
60
|
|
|
65
|
-
|
|
|
61
|
+
| 层 | 模块 | 角色 |
|
|
66
62
|
|---|---|---|
|
|
67
|
-
|
|
|
68
|
-
|
|
|
69
|
-
|
|
|
70
|
-
|
|
|
71
|
-
| **证据链** | 渲染层 | 每个数字都带**来源标签**:`[骨架:openflights]`=公开航线数据校验过"这条线能飞";`[实时API:flyai]`=刚从实时接口拉的现价现况;`[静态包:估算]`=非实时的调研估算(**下单前请核实**)。降级时标签如实更换,估算绝不冒充实时 |
|
|
63
|
+
| 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 + 每轮工具预算 + 进程护栏 |
|
|
64
|
+
| L3 | `ts/src/unified.ts` · `py/gotry_feasibility/` | 唯一求解入口(候选枚举 + 航班链 Z3) |
|
|
65
|
+
| 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)+ 实时库存桥 + OpenFlights 骨架(三值语义) |
|
|
66
|
+
| L5 | loopx 治理面 | objective / gates / evidence / quota |
|
|
72
67
|
|
|
73
|
-
|
|
68
|
+
> 完整 ADR / 演进 / 债务清单:[`docs/architecture.md`](docs/architecture.md)(英文版计划 v0.1.0)。
|
|
74
69
|
|
|
75
|
-
|
|
70
|
+
## 工具
|
|
76
71
|
|
|
77
|
-
|
|
72
|
+
六组共 21 个:
|
|
78
73
|
|
|
79
|
-
|
|
74
|
+
| 组 | 工具 | 干什么 |
|
|
75
|
+
|---|---|---|
|
|
76
|
+
| **实时检索(OTA/官方只读)** | `gotry_flyai_search` | 机票/火车/酒店实时报价(飞猪官方通道;酒店价格为上游打码展示,真实价以 jumpUrl 为准) |
|
|
77
|
+
| | `gotry_session_search` | 在**用户本人登录态**里查携程机票(授权后,只读) |
|
|
78
|
+
| | `gotry_session_login` | 登录引导:自动检测已登录与否,未登录才在用户 Chrome 弹登录入口(**零终端**) |
|
|
79
|
+
| | `gotry_weather_check` | Open-Meteo 预报≤16 天 + 历史气候基线 |
|
|
80
|
+
| | `gotry_flight_verify` | OpenSky ADS-B 航班实时观测(三值) |
|
|
81
|
+
| | `gotry_skeleton_check` | OpenFlights 168 对枢纽通航性(三值) |
|
|
82
|
+
| **库存与目录** | `gotry_hotel_search` | hotel-byte 实时桥,降级静态包(证据标注) |
|
|
83
|
+
| | `gotry_anything_search` | 城市/酒店/地标混合目录(hotel-be Anything) |
|
|
84
|
+
| **判定引擎** | `gotry_feasibility_check` | 门到门真成本可行性(Z3),逐候选判定 |
|
|
85
|
+
| **记忆与触达** | `gotry_motivation_save` | 动机画像落盘(evidence 强制,反幻觉) |
|
|
86
|
+
| | `gotry_wish_pool_add` / `gotry_wish_pool_list` | 「下一次出发」愿望池 + 0..1 条件召回 |
|
|
87
|
+
| | `gotry_companion_save` · `gotry_trip_log` | 同行人档案 / 旅行时间线 |
|
|
88
|
+
| **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 |
|
|
89
|
+
| **事实闸** | `gotry_fact_gate` | 行程产物交付前闸——见上文[工作原理](#工作原理)中的定义 |
|
|
90
|
+
| **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) |
|
|
80
91
|
|
|
81
|
-
|
|
92
|
+
## 一段对话
|
|
82
93
|
|
|
83
94
|
```
|
|
84
95
|
> 我想去洱海边发呆两三天,上海出发,预算 3000,年假别让我办公。
|
|
85
96
|
|
|
86
97
|
GoTry: 收到。先把约束记下来——
|
|
87
|
-
• 窗口: 2 天
|
|
98
|
+
• 窗口: 2 天 • 出发: 上海 • 预算: ¥3000 全含
|
|
88
99
|
• 动机: 休整 / [escape_rest: 0.7] • 还没订任何东西
|
|
89
100
|
|
|
90
101
|
引擎判定:
|
|
@@ -98,152 +109,160 @@ GoTry: 收到。先把约束记下来——
|
|
|
98
109
|
[静态包:估算] G7315/G7316 价格按 7-8 月淡季估算
|
|
99
110
|
```
|
|
100
111
|
|
|
101
|
-
> 标签导读:`[骨架:openflights]` 说的是"这条航线能飞"已被公开航线数据校验;`[实时API:*]` 说的是刚从实时接口拉回的当下数据;`[静态包:估算]`
|
|
112
|
+
> 标签导读:`[骨架:openflights]` 说的是"这条航线能飞"已被公开航线数据校验;`[实时API:*]` 说的是刚从实时接口拉回的当下数据;`[静态包:估算]` 提醒价格是淡季档估算——**订前核实**。标签由渲染层附加,模型无权染指。
|
|
102
113
|
|
|
103
|
-
|
|
114
|
+
## 快速开始
|
|
104
115
|
|
|
105
|
-
|
|
116
|
+
### npm(推荐)
|
|
106
117
|
|
|
107
|
-
|
|
118
|
+
```bash
|
|
119
|
+
npx @danceiny/gotry web
|
|
120
|
+
# → 浏览器打开 http://127.0.0.1:3080,说「我想去大理躺三天」
|
|
121
|
+
# LLM key & 模型:由 dsh 宿主 UI 配;gotry 在 CLI 层完全不出声,不要求也不回显任何凭证
|
|
122
|
+
```
|
|
108
123
|
|
|
109
|
-
|
|
|
124
|
+
| 入口 | 命令 | 什么时候用 |
|
|
110
125
|
|---|---|---|
|
|
111
|
-
|
|
|
112
|
-
| | `
|
|
113
|
-
| | `gotry_session_login` | 登录引导:自动检测已登录与否,未登录才在用户 Chrome 弹登录入口(**零终端**) |
|
|
114
|
-
| | `gotry_weather_check` | Open-Meteo 预报≤16 天 + 历史气候基线 |
|
|
115
|
-
| | `gotry_flight_verify` | OpenSky ADS-B 航班实时观测(三值) |
|
|
116
|
-
| | `gotry_skeleton_check` | OpenFlights 168 对枢纽通航性(三值) |
|
|
117
|
-
| **库存与目录** | `gotry_hotel_search` | hotel-byte 实时桥,降级静态包(证据标注) |
|
|
118
|
-
| | `gotry_anything_search` | 城市/酒店/地标混合目录(hotel-be Anything) |
|
|
119
|
-
| **判定引擎** | `gotry_feasibility_check` | 门到门真成本可行性(Z3),逐候选判定 |
|
|
120
|
-
| **记忆与触达** | `gotry_motivation_save` | 动机画像落盘(evidence 强制,反幻觉) |
|
|
121
|
-
| | `gotry_wish_pool_add` / `gotry_wish_pool_list` | 「下一次出发」愿望池 + 0..1 条件召回 |
|
|
122
|
-
| | `gotry_companion_save` · `gotry_trip_log` | 同行人档案 / 旅行时间线 |
|
|
123
|
-
| **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 |
|
|
124
|
-
| **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」 |
|
|
125
|
-
| **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) |
|
|
126
|
+
| Web 对话(推荐) | `npx @danceiny/gotry web` | 持续多轮规划,看推理可视化 → `:3080` |
|
|
127
|
+
| headless 一问一答 | `npx @danceiny/gotry "我想从深圳休整两天,预算 3000"` | 脚本 / CI / 定向调试 → stdout |
|
|
126
128
|
|
|
127
|
-
|
|
129
|
+
前置:Node ≥ 22.15。LLM 凭证由 dsh 宿主 UI 配,OpenAI 兼容端点(MiniMax/中转/自建网关)走 dsh 的模型设置。首启 6–15 秒属正常冷启动;`:3080` 被占先腾端口;异常退出会留证据到 `gotry-state/incidents.jsonl`(不静默)。
|
|
128
130
|
|
|
129
|
-
|
|
131
|
+
> **成本核算** —— `ts/data/llm-price-table.json`(schema `gotry_llm_price_table_v2`)是 nightly 成本核算的唯一事实源。新增模型或换中转=对该文件提 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;`--fetch` 拉官方页)。**永不自动 apply 价格**。
|
|
130
132
|
|
|
131
|
-
|
|
133
|
+
### 开发者源码安装
|
|
132
134
|
|
|
133
|
-
|
|
134
|
-
|
|
135
|
-
|
|
136
|
-
|
|
135
|
+
```bash
|
|
136
|
+
git clone https://github.com/Danceiny/gotry && cd gotry
|
|
137
|
+
npm ci && npm --prefix ts ci # 锁定的 root/TS 依赖闭包
|
|
138
|
+
node scripts/build-dist.mjs # 构建 JS runtime
|
|
139
|
+
./gotry web # 仓内入口,与 npm 形态同 UX
|
|
140
|
+
```
|
|
137
141
|
|
|
138
|
-
|
|
142
|
+
源码入口与 npm 包解析同一组 216 个精确直接依赖的 DSH `0.1.2-alpha.3` closure(publish preverify 拒绝漏钉、混版和 range)。源码普通运行状态落在 `ts/dsh-runtime/gotry-state/`;benchmark opt-in 与 npm 包运行用调用目录隔离。
|
|
139
143
|
|
|
140
|
-
|
|
144
|
+
## 账号会话:授权与隐私
|
|
141
145
|
|
|
142
|
-
|
|
146
|
+
账号会话通道用**你本人已登录的 Chrome**读酒店/机票实时数据,为此立了四条 hard 规则:
|
|
143
147
|
|
|
144
|
-
|
|
145
|
-
|
|
146
|
-
|
|
147
|
-
|
|
148
|
-
│ L3 领域 统一行程模型 + Z3 可行性引擎(枚举/Z3 双形态) │
|
|
149
|
-
│ L4 数据 静态数据包 + hotelbyte-cli 实时桥 + OpenFlights 骨架 │
|
|
150
|
-
│ L5 治理 LoopX(objective / gates / evidence / quota) │
|
|
151
|
-
└──────────────────────────────────────────────────────────────┘
|
|
152
|
-
```
|
|
148
|
+
1. **登录在外部网站完成** —— gotry 从不提供、不代填、不收集任何密码/验证码/cookie 值。它只回答一个布尔问题:"登录票据 cookie 存在吗"(只读**名字**,0 值过手)。已登录会被自动识别,零弹窗。
|
|
149
|
+
2. **授权卡,每会话一次** —— 首次动用账号会话弹运行时审批卡;批准后会话内记住,拒绝即本会话吊销,不再打扰。总闸 `sessionAccess: ask|allow|off` 随时可关。
|
|
150
|
+
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,遇验证码立即停、交还给你。
|
|
151
|
+
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。
|
|
153
152
|
|
|
154
|
-
|
|
155
|
-
|---|---|---|
|
|
156
|
-
| 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 |
|
|
157
|
-
| L3 | `ts/src/unified.ts` · `py/gotry_feasibility/` | 唯一求解入口(候选枚举 + 航班链 Z3) |
|
|
158
|
-
| 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) |
|
|
159
|
-
| L5 | loopx 治理面 | objective / gates / evidence / quota |
|
|
153
|
+
前置(一次性):[GoTry Session Bridge](https://chromewebstore.google.com/detail/gotry-session-bridge/oeajpiccmonococjcegddlooeeohlbgd) Chrome 应用商店一键装(自动更新)。账号会话工具首次需要扩展时,**由 dsh 宿主 UI 把商店 URL 作为可点链接渲染**;gotry 这边不会跑任何 setup wizard、也不会让你手动加载已解压扩展。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);后台 health-watch 探活,扩展一就位自动重放你的检索。未安装时工具返回 `needs-extension` 并把商店链接置于 verdict,不消耗执行配额。
|
|
160
154
|
|
|
161
|
-
|
|
155
|
+
## 构造上可信
|
|
162
156
|
|
|
163
|
-
|
|
157
|
+
1. **模型只翻译,求解器才判决** —— LLM 永远不产出可行性判决与算术;那些由 Z3 基于抽取事实计算。
|
|
158
|
+
2. **每个数字带来源标签** —— 由渲染层附加,模型无权染指;降级时如实更换,估算绝不冒充实时。
|
|
159
|
+
3. **不存在写路径** —— 预订/支付类工具必须先过 WriteGate 才允许实现;未来的预订缝已被 `booking_saga_fsm.v1` 边表钉住。
|
|
160
|
+
4. **登录永不碰凭证** —— 登录发生在外部网站;gotry 只读 cookie 名;授权每会话问一次、可吊销。
|
|
161
|
+
5. **检索物理只读** —— ReadGuard 在网络层中止写请求;验证码让 agent 停下、把控制权交还给你。
|
|
162
|
+
6. **回溯不到就是 blocked** —— 事实闸拒绝交付任何可下单 claim 无法回溯到 exact-date 工具结果的行程——绝不宣称「已验证方案」。
|
|
163
|
+
7. **价格 fail-closed** —— 未知模型不猜价;价表只经 PR 变更;漂移监测只报告、永不自动 apply。
|
|
164
|
+
8. **你的数据是你的** —— 产品状态在 `gotry-state/`;自动化测试与 smoke 用隔离 state root,永不写创始人的真实产品数据。
|
|
165
|
+
|
|
166
|
+
## 状态与限制
|
|
164
167
|
|
|
165
|
-
|
|
168
|
+
当前版本:**v0.0.1-rc.17**(npm dist-tag `rc`;`latest` 仍指 rc.16,沿用 dsh-管-LLM 安装口径)。评测处于 Phase 0 基座——确定性合同、校验器与节奏策略;无外部 benchmark 分数、无花费、无 uplift 声明。
|
|
166
169
|
|
|
167
|
-
|
|
170
|
+
**今天可用**(全栈回归全绿;每项都有确定性测试):
|
|
168
171
|
|
|
169
|
-
- **Z3 求解引擎** ——
|
|
170
|
-
-
|
|
171
|
-
-
|
|
172
|
-
-
|
|
173
|
-
-
|
|
174
|
-
-
|
|
175
|
-
- **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)。
|
|
176
|
-
- **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`)
|
|
172
|
+
- **Z3 求解引擎** —— 可行性判决 + 门到门全成本;历史并发竞态已根治并进回归闸
|
|
173
|
+
- **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;实时票价可覆写求解价(`GOTRY_REALTIME_PRICING=1`)
|
|
174
|
+
- **账号会话检索** —— 你本人登录态查携程机票;观测轮次中所有可评分 hit 全过、ReadGuard 零写,非 hit 保持显式 `miss` 记录——不作超出此口径的实时可售声明
|
|
175
|
+
- **扩展按需装** —— `[GoTry Session Bridge](https://chromewebstore.google.com/detail/gotry-session-bridge/oeajpiccmonococjcegddlooeeohlbgd)` 由 dsh 宿主 UI 在账号会话工具首次需要时以可点链接给出(Chrome 商店一键装 + 自动更新);gotry 这边不跑 setup wizard、不开 chrome://extensions、不动剪贴板
|
|
176
|
+
- **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`)
|
|
177
|
+
- **有界的工具循环** —— 16 次派发后注入软收敛,18 次后结构化 `TOOL_BUDGET_EXHAUSTED` 拒绝;经打包消费者安装的 E2E 在 CI 里实测
|
|
177
178
|
|
|
178
|
-
**已知限制**(
|
|
179
|
+
**已知限制**(诚实清单):
|
|
179
180
|
|
|
180
|
-
-
|
|
181
|
-
-
|
|
182
|
-
-
|
|
183
|
-
-
|
|
184
|
-
-
|
|
185
|
-
- ⏳ **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。
|
|
181
|
+
- **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
|
|
182
|
+
- **酒店会话适配** —— 携程酒店/美团登录态面等实测回填;机票已通
|
|
183
|
+
- **界面语言** —— 英文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
|
|
184
|
+
- **外部 benchmark 泛化** —— 迄今所有冻结外部运行均仅 diagnostic(无分数、无 uplift 声明);逐轮工程台账见 [`docs/benchmark-environment-bridge.md`](docs/benchmark-environment-bridge.md)
|
|
185
|
+
- **预订** —— 今天没有任何可下单路径;M5 只经 WriteGate 与 booking-saga 状态机启封
|
|
186
186
|
|
|
187
187
|
<details>
|
|
188
|
-
<summary
|
|
188
|
+
<summary>更深的工程状态(账本合同 / 证据合同 / 里程碑口径)</summary>
|
|
189
189
|
|
|
190
|
-
|
|
190
|
+
状态权威面在文档,不在 README:事务化状态账本(ADR-15)+ 双形态冻结(ADR-16:本地+Web 一套账本语义);M3 真实 cohort 证据合同已立(fixture 不充当 Exit,真实 50–200 人样本即开 Exit);M4 paired-cohort 价值证据合同(合成数据不充当 Exit 证据);异步工单终态合同(`gotry_async_terminal.v1`:4/4 → succeeded / ledger settled / exit 0)。细则见 [`docs/roadmap.md`](docs/roadmap.md) / [`docs/architecture.md`](docs/architecture.md) §1 与 #19–#22。
|
|
191
191
|
|
|
192
192
|
</details>
|
|
193
193
|
|
|
194
|
-
|
|
195
|
-
|
|
196
|
-
## 📚 文档 — Docs
|
|
194
|
+
## 路线图
|
|
197
195
|
|
|
198
|
-
|
|
|
199
|
-
|
|
200
|
-
|
|
|
201
|
-
|
|
|
202
|
-
|
|
|
203
|
-
|
|
|
204
|
-
|
|
|
205
|
-
|
|
|
206
|
-
|
|
|
207
|
-
| [`docs/release-notes.md`](docs/release-notes.md) | 历史:每个 tag 的发布闸勾稽 |
|
|
196
|
+
| # | 里程碑 | 范围 | 状态 |
|
|
197
|
+
|---|---|---|---|
|
|
198
|
+
| M0 | 确定性管道 | 引擎双实现 + 真实数据包 + 对账框架 | ✅ |
|
|
199
|
+
| M1 | Agent 形态成立 | LLM 进环;对话即界面;gates 选择题 | ✅ 2026-08-22 |
|
|
200
|
+
| M2 | 实时数据 | hotelbyte 桥 + 航班源;证据链换实时标签 | ✅ 2026-08-22 |
|
|
201
|
+
| M3 | 最小可用产品 | 最小 Web 面 + 50–200 种子用户(洱海/普吉场景) | **← 当前 —— evidence 未收口** |
|
|
202
|
+
| M4 | 记忆与「下一次出发」 | 六层记忆 C 端域;paired-cohort 价值证据 | founder 授权并行 |
|
|
203
|
+
| M5 | 交易闭环 | WriteGate 上生产;预订 / 支付 / 退改 | entry gate 启封 |
|
|
204
|
+
| M6 | B2B 包裹 | principal/sponsor 插件,内核零改动 | entry gate 启封 |
|
|
208
205
|
|
|
209
|
-
|
|
206
|
+
唯一权威时间线(逐里程碑的进入/退出条件、交付物与 gate):[`docs/roadmap.md`](docs/roadmap.md)。
|
|
210
207
|
|
|
211
|
-
##
|
|
208
|
+
## 跑测试
|
|
212
209
|
|
|
213
210
|
```bash
|
|
214
|
-
./scripts/run-all-tests.sh
|
|
211
|
+
./scripts/run-all-tests.sh # 全栈套件(纯 TS,无 Python 依赖)
|
|
212
|
+
cd ts
|
|
213
|
+
npx tsx scripts/evaluation-contract-tests.ts # 评测 Phase 0 合同(离线)
|
|
214
|
+
npx tsx scripts/evaluation-cadence-tests.ts # 确定性节奏策略/planner
|
|
215
215
|
```
|
|
216
216
|
|
|
217
|
-
|
|
217
|
+
套件覆盖金标准引擎、对话重放、跨进程异步工单、插件 smoke、实时桥、进程护栏、i18n、记忆域、Z3 并发闸、事实闸、打包消费者工具预算 E2E 等;权威分节以 `scripts/run-all-tests.sh` 实际枚举为准。真实会话 benchmark(`npx tsx scripts/sf-live-benchmark.ts --golden=static`)为 opt-in,需你的 Chrome 会话扩展在线,永不进 CI。
|
|
218
218
|
|
|
219
|
-
|
|
219
|
+
## 参与开发
|
|
220
220
|
|
|
221
|
-
|
|
221
|
+
从最新 `main` 切出 `feat/ · fix/ · docs/ · chore/` 分支,本地全栈绿后开 Pull Request——`main` 不直接推。CI(Node 22/24,typecheck + 全部套件)与维护者 review 双绿后 squash 合入。**测试红着不许合。** 完整指南:[CONTRIBUTING.md](CONTRIBUTING.md)。Bug/功能建议:用 issue 模板(先搜既有 issue)。
|
|
222
222
|
|
|
223
|
-
|
|
223
|
+
## 给 AI Agent
|
|
224
224
|
|
|
225
|
-
|
|
225
|
+
如果你是在本仓工作的 agent,[`AGENTS.md`](AGENTS.md) 是绑定契约,先读它。要点:
|
|
226
226
|
|
|
227
|
-
-
|
|
228
|
-
-
|
|
229
|
-
-
|
|
227
|
+
- **入场先清扫异步工单**:`ts/gotry-state/async/*.json` 无同名 `.deliverable.md` → `cd ts && npx tsx scripts/async-collect.ts <id>`。
|
|
228
|
+
- **分层纪律**:算术只在 `model.ts` / `unified.py` 的 evaluate 层;求解只在 `unified.ts` / `unified.py`;`engine.*` / `journey.*` 是 deprecated 兼容层,新代码不得调用。改任何一侧必须跑全栈回归。
|
|
229
|
+
- **绝不写共享状态**:`ts/dsh-runtime/gotry-state/` 是创始人真实产品数据;验证写路径只用隔离 `stateRoot`。
|
|
230
|
+
- **状态同步纪律**:任何改变系统形态/状态/债务的提交,必须在同一提交内同步 `architecture.md` §11 六状态面;只暂存具名文件——禁止 `git add -A`。
|
|
230
231
|
|
|
231
|
-
|
|
232
|
+
程序层语境:[`docs/gotry-master-outline.md`](docs/gotry-master-outline.md)。技术权威面:[`docs/architecture.md`](docs/architecture.md)。
|
|
232
233
|
|
|
233
|
-
##
|
|
234
|
+
## 文档
|
|
234
235
|
|
|
235
|
-
|
|
236
|
+
| 文档 | 内容 |
|
|
237
|
+
|---|---|
|
|
238
|
+
| [`docs/architecture.md`](docs/architecture.md) | 系统 / ADR / 演进 / 债务清单(中文,权威) |
|
|
239
|
+
| [`docs/gotry-master-outline.md`](docs/gotry-master-outline.md) | 总纲:工作分解 · 复用矩阵 |
|
|
240
|
+
| [`docs/gotry-product-design.md`](docs/gotry-product-design.md) | 产品设计:主循环 · 透明机制 · 全成本模型 |
|
|
241
|
+
| [`docs/roadmap.md`](docs/roadmap.md) | M0–M6 时间线与当前位置 |
|
|
242
|
+
| [`docs/user-guide.md`](docs/user-guide.md) | 终端用户使用指南 |
|
|
243
|
+
| [`docs/data-sources.md`](docs/data-sources.md) | 数据源与证据链政策 |
|
|
244
|
+
| [`docs/extension-privacy.md`](docs/extension-privacy.md) | Session Bridge 扩展隐私 |
|
|
245
|
+
| [`docs/benchmark-environment-bridge.md`](docs/benchmark-environment-bridge.md) | 外部 benchmark 桥——工程台账 |
|
|
246
|
+
| [`docs/evaluation-foundation.md`](docs/evaluation-foundation.md) | 评测 Phase 0 基座 |
|
|
247
|
+
| [`docs/booking-saga-fsm.md`](docs/booking-saga-fsm.md) | 预订 saga 状态机(M5 缝词汇) |
|
|
248
|
+
| [`docs/kimi-postmortem.md`](docs/kimi-postmortem.md) | 一次真实 AI 旅行规划失败复盘(反面教材) |
|
|
249
|
+
| [`docs/release-notes.md`](docs/release-notes.md) | 逐版本发布决策(「为什么」) |
|
|
250
|
+
| [`CHANGELOG.md`](CHANGELOG.md) | 机器衍生的变更日志(Keep a Changelog + Conventional Commits) |
|
|
251
|
+
| [`docs/tokens.md`](docs/tokens.md) | npm 2FA / 发布机制 |
|
|
252
|
+
|
|
253
|
+
## License
|
|
254
|
+
|
|
255
|
+
**MIT**——与上游 dsh 一致。文本见 [LICENSE](LICENSE)。
|
|
236
256
|
|
|
237
257
|
---
|
|
238
258
|
|
|
239
|
-
|
|
259
|
+
**Built with**: DeepSeek Harness 0.1.2-alpha.3 (root-pinned) · Cordis · Z3 (WASM) · loopx (pipx) · hotelbyte-cli · Agent-Reach v1.5.0 · OpenFlights · TypeScript
|
|
240
260
|
|
|
241
|
-
-
|
|
242
|
-
- 深 README(中文)面向中国出境首发种子用户群;英文版是它的完整镜像(不含深度工程文档)。
|
|
243
|
-
- `docs/` 深度工程文档当前中文先行,英文版计划 v0.1.0 同步。
|
|
261
|
+
**版本基线:`v0.0.1-rc.17`(npm dist-tag `rc`;`latest` 沿用 rc.16)。** 当前 checkout 的权威验证闸以 `scripts/run-all-tests.sh` 实际枚举为准;发布流程见 `scripts/publish-npm.sh`。
|
|
244
262
|
|
|
245
263
|
---
|
|
246
264
|
|
|
247
|
-
|
|
265
|
+
## 中英版说明
|
|
248
266
|
|
|
249
|
-
|
|
267
|
+
- 本文件与 [README.md](README.md) 各自完整自含、结构互为镜像(常见开源双语布局)。
|
|
268
|
+
- `docs/` 深度工程文档当前中文先行,英文版计划 v0.1.0 同步。
|
package/bin/gotry-bootstrap.js
CHANGED
|
@@ -440,12 +440,10 @@ async function runInlineHealthWatch(timeoutMs) {
|
|
|
440
440
|
}
|
|
441
441
|
|
|
442
442
|
async function main() {
|
|
443
|
-
|
|
444
|
-
|
|
445
|
-
|
|
446
|
-
|
|
447
|
-
// → dsh UI 渲商店 URL → 用户点链接 → 装好 → 同会话内自动 retry(health-watch)。
|
|
448
|
-
if (WIZARD) {
|
|
443
|
+
// wizard 子命令(2026-09-02 商店上架后退化):**只走 stdout 提示 + 健康探活等待**;
|
|
444
|
+
// 不 spawn 任何 GUI 工具(不动 pbcopy / osascript / open / xdg-open / zenity),
|
|
445
|
+
// 不打开 chrome://extensions,不动扩展路径——浏览器自己当安装器,gotry 不越界。
|
|
446
|
+
if (WIZARD) {
|
|
449
447
|
if (WIZARD_DRY_RUN) {
|
|
450
448
|
say('[gotry-wizard] dry-run(零网络零浏览器零剪贴板;run-all §40 走这条)')
|
|
451
449
|
say(' 步骤: ensure-extension-files → watch-extension-ready(2 步纯 Node 端)')
|
|
@@ -467,44 +465,33 @@ async function main() {
|
|
|
467
465
|
const watchResult = await runHealthWatch()
|
|
468
466
|
if (watchResult.ready) {
|
|
469
467
|
say(`[gotry-wizard] ✅ 扩展就绪(${watchResult.attempts} 次探活,等待 ${watchResult.waitedMs}ms)`)
|
|
470
|
-
say('[gotry-wizard] 现在 `npx gotry web` → dsh UI → 调 gotry_session_search 即可拿到结果。')
|
|
471
468
|
process.exit(0)
|
|
472
469
|
} else {
|
|
473
470
|
say(`[gotry-wizard] ✗ ${watchResult.reason}——未在 ${watchResult.timeoutMs}ms 内就绪`)
|
|
474
|
-
say('[gotry-wizard] dsh UI 中
|
|
471
|
+
say('[gotry-wizard] 不需重跑;扩展未装时 dsh UI 中 gotry_session_search 的 needs-extension 会带商店链接,wizard 完全幂等。')
|
|
475
472
|
process.exit(1)
|
|
476
473
|
}
|
|
477
474
|
}
|
|
478
475
|
|
|
479
476
|
if (process.platform === 'win32') {
|
|
480
|
-
say('[gotry-setup]
|
|
481
|
-
say(` hbcli: ${HBCLI_INSTALL_CMD}(WSL);agent-reach: python -m venv .venv && .venv/Scripts/pip install ${REACH_INSTALL_URL}`)
|
|
482
|
-
say(' dsh-better-sidebar: npx -y --package @deepseek-ai/dsh dsh plugin --profile web add dsh-better-sidebar@latest')
|
|
483
|
-
say(' GoTry Session Bridge 扩展:推荐 Chrome 应用商店一键安装 https://chromewebstore.google.com/detail/gotry-session-bridge/oeajpiccmonococjcegddlooeeohlbgd ;本地通道:手动把包内 extension/ 目录拷到 %USERPROFILE%\\.gotry\\extension,再在 chrome://extensions 开发者模式「加载已解压的扩展程序」')
|
|
477
|
+
say('[gotry-setup] GoTry Session Bridge 扩展:推荐 Chrome 应用商店一键安装 https://chromewebstore.google.com/detail/gotry-session-bridge/oeajpiccmonococjcegddlooeeohlbgd ;本地通道:手动把包内 extension/ 目录拷到 %USERPROFILE%\\.gotry\\extension,再在 chrome://extensions 开发者模式「加载已解压的扩展程序」')
|
|
484
478
|
process.exit(AUTO ? 0 : 1)
|
|
485
479
|
}
|
|
486
480
|
if (AUTO && (process.env.CI || process.env.GOTRY_SETUP_SKIP === '1')) {
|
|
487
|
-
say('[gotry-setup] CI/GOTRY_SETUP_SKIP
|
|
488
|
-
say('GoTry installed. Run: npx gotry web (dsh Web UI on :3080)')
|
|
481
|
+
say('[gotry-setup] CI/GOTRY_SETUP_SKIP 检测——跳过')
|
|
489
482
|
process.exit(0)
|
|
490
483
|
}
|
|
491
484
|
if (!AUTO && process.env.GOTRY_SETUP_SKIP === '1') { say('[gotry-setup] GOTRY_SETUP_SKIP=1——跳过'); process.exit(0) }
|
|
492
485
|
const results = []
|
|
493
|
-
|
|
494
|
-
else say('[gotry-setup] hbcli:GOTRY_SETUP_HBCLI=0 跳过')
|
|
495
|
-
if (process.env.GOTRY_SETUP_REACH !== '0') results.push(await setupReach())
|
|
496
|
-
else say('[gotry-setup] agent-reach:GOTRY_SETUP_REACH=0 跳过')
|
|
497
|
-
if (process.env.GOTRY_SETUP_SIDEBAR !== '0') results.push(await setupSidebar())
|
|
498
|
-
else say('[gotry-setup] dsh-better-sidebar:GOTRY_SETUP_SIDEBAR=0 跳过')
|
|
486
|
+
// gotry 自留面只剩扩展是否就位;hbcli/agent-reach/dsh-better-sidebar 由各自宿主生态自管。
|
|
499
487
|
if (process.env.GOTRY_SETUP_EXTENSION !== '0') results.push(await (EXTENSION_FROM === 'github' ? setupExtensionFromGithub() : setupExtension()))
|
|
500
488
|
else say('[gotry-setup] GoTry Session Bridge 扩展:GOTRY_SETUP_EXTENSION=0 跳过')
|
|
501
|
-
say('[gotry-setup] flyai:无需安装(npx 每次自拉 @fly-ai/flyai-cli,免 key)')
|
|
502
489
|
const failed = results.filter((r) => !r.ok).length
|
|
503
490
|
if (failed > 0) {
|
|
504
|
-
say(`[gotry-setup] ${failed} 项未就绪——gotry
|
|
491
|
+
say(`[gotry-setup] ${failed} 项未就绪——gotry 本体不受影响;可稍后重跑: npx gotry setup`)
|
|
505
492
|
process.exit(AUTO ? 0 : 1)
|
|
506
493
|
}
|
|
507
|
-
say('[gotry-setup]
|
|
494
|
+
say('[gotry-setup] 扩展就绪检查完成。')
|
|
508
495
|
process.exit(0)
|
|
509
496
|
}
|
|
510
497
|
|
package/bin/gotry-inner.js
CHANGED
|
@@ -86,10 +86,6 @@ Usage:
|
|
|
86
86
|
gotry "一段完整任务..." # headless 一问一答
|
|
87
87
|
gotry help # this help
|
|
88
88
|
|
|
89
|
-
Prerequisites:
|
|
90
|
-
• Node 22.15+
|
|
91
|
-
• \`.env\` 里 LLM_API_KEY (DeepSeek / OpenAI 兼容均可;自定义端点另配 LLM_BASE_URL,一般以 /v1 结尾;指定模型另配 LLM_MODEL,同时驱动 dsh 会话面与仓内脚本)
|
|
92
|
-
|
|
93
89
|
Detail: https://github.com/Danceiny/gotry — README
|
|
94
90
|
`)
|
|
95
91
|
process.exit(0)
|
|
@@ -450,20 +446,8 @@ const terminateOnSignal = (signal, listener) => {
|
|
|
450
446
|
process.off(signal, listener)
|
|
451
447
|
process.kill(process.pid, signal)
|
|
452
448
|
}
|
|
453
|
-
|
|
454
|
-
|
|
455
|
-
process.once('SIGINT', onSigint)
|
|
456
|
-
process.once('SIGTERM', onSigterm)
|
|
457
|
-
if (!process.env.DEEPSEEK_API_KEY && mode !== 'help') {
|
|
458
|
-
console.error('[gotry] 缺少 LLM API key —— 两种方式任选其一后重跑:')
|
|
459
|
-
console.error(` 1) 在当前目录创建 .env 写入一行: LLM_API_KEY=<你的 DeepSeek key>(key 从 https://platform.deepseek.com 获取)`)
|
|
460
|
-
console.error(' 2) 或临时环境变量: export LLM_API_KEY=<key>')
|
|
461
|
-
cleanupPatch()
|
|
462
|
-
process.exit(1)
|
|
463
|
-
}
|
|
464
|
-
|
|
465
|
-
// --- 调 vendored dsh 二进制(不走 npx)---
|
|
466
|
-
// headless: 第一个非 -- 之后的参数是 task —— gotry argv 0 是 headless 触发,其余都算 task;
|
|
449
|
+
// LLM key 由 dsh 宿主管理(凭证是用户资产,UI 在 dsh 里;gotry 不拦截启动期,
|
|
450
|
+
// 不在用户面前展示任何 key 配置引导),此处直接放手 spawn dsh。
|
|
467
451
|
const binJs = mode === 'web'
|
|
468
452
|
? ['web', '--patch', patchPath, ...(process.argv.includes('--no-open') ? ['--no-open'] : [])]
|
|
469
453
|
: ['--profile', 'headless', '--patch', patchPath, ...rest]
|
|
@@ -32,20 +32,19 @@ const c1 = runBootstrap([
|
|
|
32
32
|
'--check-only'
|
|
33
33
|
], {});
|
|
34
34
|
assert.equal(c1.code, 0, `--check-only 应 exit 0,实际 ${c1.code}\n${c1.out}`);
|
|
35
|
-
assert.ok(c1.out.includes('
|
|
36
|
-
assert.ok(c1.out.includes('
|
|
37
|
-
assert.ok(c1.out.includes('
|
|
38
|
-
assert.ok(c1.out.includes('dsh-better-sidebar'), '
|
|
39
|
-
|
|
40
|
-
console.log('1. --check-only 探测报告 exit 0(hbcli/agent-reach/flyai/dsh-better-sidebar/会话扩展 五节齐)OK');
|
|
35
|
+
assert.ok(c1.out.includes('Session Bridge'), '报告应含扩展就位节(issue #21 传输层方案 C)');
|
|
36
|
+
assert.ok(!c1.out.includes('hbcli'), 'setup 不再替用户管 hbcli(已让出)');
|
|
37
|
+
assert.ok(!c1.out.includes('agent-reach'), 'setup 不再替用户管 agent-reach(已让出)');
|
|
38
|
+
assert.ok(!c1.out.includes('dsh-better-sidebar'), 'setup 不再替用户管 dsh-better-sidebar(已让出)');
|
|
39
|
+
console.log('1. --check-only 探测报告 exit 0(扩展就位 + 让出 hbcli/agent-reach/sidebar)OK');
|
|
41
40
|
const c2 = runBootstrap([
|
|
42
41
|
'--auto'
|
|
43
42
|
], {
|
|
44
43
|
GOTRY_SETUP_SKIP: '1'
|
|
45
44
|
});
|
|
46
|
-
assert.equal(c2.code, 0, '--auto 跳过态应 exit 0
|
|
45
|
+
assert.equal(c2.code, 0, '--auto 跳过态应 exit 0');
|
|
47
46
|
assert.ok(c2.out.includes('跳过'), '应输出跳过说明');
|
|
48
|
-
console.log('2. --auto + GOTRY_SETUP_SKIP=1 → exit 0(
|
|
47
|
+
console.log('2. --auto + GOTRY_SETUP_SKIP=1 → exit 0(永不挡 npm install)OK');
|
|
49
48
|
const c3 = runBootstrap([], {
|
|
50
49
|
GOTRY_SETUP_SKIP: '1'
|
|
51
50
|
});
|
|
@@ -55,45 +54,32 @@ console.log('3. 显式模式 + GOTRY_SETUP_SKIP=1 → exit 0 OK');
|
|
|
55
54
|
const c4 = runBootstrap([
|
|
56
55
|
'--check-only'
|
|
57
56
|
], {
|
|
58
|
-
|
|
57
|
+
GOTRY_SETUP_EXTENSION: '0'
|
|
59
58
|
});
|
|
60
59
|
assert.equal(c4.code, 0, '单项跳过态应 exit 0');
|
|
61
|
-
assert.ok(c4.out.includes('
|
|
62
|
-
|
|
63
|
-
console.log('4. GOTRY_SETUP_SIDEBAR=0 单项跳过(其余节照常)OK');
|
|
60
|
+
assert.ok(c4.out.includes('GOTRY_SETUP_EXTENSION=0 跳过'), '应输出扩展单项跳过说明');
|
|
61
|
+
console.log('4. GOTRY_SETUP_EXTENSION=0 单项跳过 OK');
|
|
64
62
|
const c5 = runBootstrap([
|
|
65
|
-
'--check-only'
|
|
66
|
-
], {});
|
|
67
|
-
assert.ok(c5.out.includes('.gotry/extension'), '扩展报告应含落位路径 ~/.gotry/extension(绝对路径形态)');
|
|
68
|
-
const c5b = runBootstrap([
|
|
69
|
-
'--check-only'
|
|
70
|
-
], {
|
|
71
|
-
GOTRY_SETUP_EXTENSION: '0'
|
|
72
|
-
});
|
|
73
|
-
assert.equal(c5b.code, 0, '扩展单项跳过态应 exit 0');
|
|
74
|
-
assert.ok(c5b.out.includes('GOTRY_SETUP_EXTENSION=0 跳过'), '应输出扩展单项跳过说明');
|
|
75
|
-
console.log('5. 会话扩展节(check-only 报告 + GOTRY_SETUP_EXTENSION=0 单项跳过)OK');
|
|
76
|
-
const c6 = runBootstrap([
|
|
77
63
|
'wizard',
|
|
78
64
|
'--dry-run'
|
|
79
65
|
], {});
|
|
80
|
-
assert.equal(
|
|
81
|
-
assert.ok(
|
|
82
|
-
assert.ok(
|
|
83
|
-
assert.ok(
|
|
84
|
-
assert.ok(
|
|
85
|
-
console.log('
|
|
86
|
-
const
|
|
66
|
+
assert.equal(c5.code, 0, `wizard --dry-run 应 exit 0,实际 ${c5.code}\n${c5.out}`);
|
|
67
|
+
assert.ok(c5.out.includes('dry-run'), '应输出 dry-run 字样');
|
|
68
|
+
assert.ok(c5.out.includes('gotry-wizard'), '应输出 [gotry-wizard] 标签');
|
|
69
|
+
assert.ok(c5.out.includes('ensure-extension-files'), 'dry-run 应列 ensure-extension-files');
|
|
70
|
+
assert.ok(c5.out.includes('watch-extension-ready'), 'dry-run 应列 watch-extension-ready');
|
|
71
|
+
console.log('5. wizard 子命令(--dry-run 零网络,2 步齐全 + 极简 stdout)OK');
|
|
72
|
+
const c6 = runBootstrap([
|
|
87
73
|
'wizard'
|
|
88
74
|
], {
|
|
89
75
|
GOTRY_SETUP_EXTENSION: '0',
|
|
90
76
|
GOTRY_ONBOARDING_TIMEOUT_MS: '600',
|
|
91
77
|
GOTRY_ONBOARDING_INTERVAL_MS: '200'
|
|
92
78
|
});
|
|
93
|
-
assert.equal(
|
|
94
|
-
assert.ok(
|
|
95
|
-
console.log('
|
|
96
|
-
const
|
|
79
|
+
assert.equal(c6.code, 1, `wizard(超时)应 exit 1,实际 ${c6.code}\n${c6.out}`);
|
|
80
|
+
assert.ok(c6.out.includes('gotry-wizard'), '应输出 [gotry-wizard] 标签');
|
|
81
|
+
console.log('6. wizard 真实路径(扩展未就绪,exit 1 + 心跳)OK');
|
|
82
|
+
const c7 = runBootstrap([
|
|
97
83
|
'--check-only',
|
|
98
84
|
'--extension-from=github'
|
|
99
85
|
], {
|
|
@@ -102,10 +88,10 @@ const c8 = runBootstrap([
|
|
|
102
88
|
GOTRY_SETUP_SIDEBAR: '0',
|
|
103
89
|
GOTRY_EXTENSION_RELEASE_BASE: 'http://127.0.0.1:1/releases'
|
|
104
90
|
});
|
|
105
|
-
assert.equal(
|
|
106
|
-
assert.ok(
|
|
107
|
-
assert.ok(
|
|
108
|
-
const
|
|
91
|
+
assert.equal(c7.code, 0, `github 通道降级应 exit 0,实际 ${c7.code}\n${c7.out}`);
|
|
92
|
+
assert.ok(c7.out.includes('GitHub Releases 下载通道'), '应打印 github 通道标题');
|
|
93
|
+
assert.ok(c7.out.includes('降级包内副本'), '失败应显式降级 bundled');
|
|
94
|
+
const c7b = runBootstrap([
|
|
109
95
|
'--check-only',
|
|
110
96
|
'--extension-from=不合法值'
|
|
111
97
|
], {
|
|
@@ -113,10 +99,10 @@ const c8b = runBootstrap([
|
|
|
113
99
|
GOTRY_SETUP_REACH: '0',
|
|
114
100
|
GOTRY_SETUP_SIDEBAR: '0'
|
|
115
101
|
});
|
|
116
|
-
assert.equal(
|
|
117
|
-
assert.ok(!
|
|
118
|
-
console.log('
|
|
119
|
-
console.log('BOOTSTRAP TESTS:
|
|
102
|
+
assert.equal(c7b.code, 0, `非法 --extension-from 值应回落 bundled 且 exit 0,实际 ${c7b.code}\n${c7b.out}`);
|
|
103
|
+
assert.ok(!c7b.out.includes('下载通道'), '非法值不应进入 github 通道');
|
|
104
|
+
console.log('7. 扩展分发 github 通道(不可达基址即时降级 + 非法值回落 bundled)OK');
|
|
105
|
+
console.log('BOOTSTRAP TESTS: 7/7 OK(扩展就位 + 跳过开关 / wizard --dry-run / wizard 真实 / 扩展分发通道 / 显式跳过 + auto 跳过 + 单项跳过)');
|
|
120
106
|
|
|
121
107
|
|
|
122
108
|
//# sourceURL=ts/scripts/bootstrap-tests.ts
|