flow-walker-cli 0.2.0

package/AGENTS.md

@@ -0,0 +1,299 @@
+# flow-walker — Agent Workflow Guide
+
+## What is flow-walker
+
+flow-walker is the **flow layer** — it discovers, executes, and reports on app flows.
+It uses [agent-flutter](https://github.com/beastoin/agent-flutter) and [agent-swift](https://github.com/beastoin/agent-swift) as **transport layers** that control specific platforms.
+
+**Six commands:**
+- `walk` — BFS-explore the app, discover screens, generate YAML flows
+- `run` — Execute a YAML flow, produce run.json + video + screenshots
+- `report` — Generate self-contained HTML report from run results
+- `push` — Upload report to hosted service, return shareable URL
+- `get` — Fetch run data from hosted service by run ID
+- `schema` — Machine-readable command introspection (agent discovery)
+
+## Agent-first workflow
+
+```bash
+# 1. Discover available commands
+flow-walker schema                    # → { version, commands: [...] }
+flow-walker schema run                # → args, flags with types, exit codes
+
+# 2. Dry-run to verify flow resolves
+flow-walker run flow.yaml --dry-run   # → per-step resolved/unresolved + reasons
+
+# 3. Execute
+flow-walker run flow.yaml --json      # → run.json with unique run ID
+
+# 4. Report
+flow-walker report ./run-output/<run-id>/
+
+# 5. Share (hosted)
+flow-walker push ./run-output/<run-id>/ --json  # → { id, url, htmlUrl, expiresAt }
+
+# 6. Retrieve run data later
+flow-walker get 25h7afGwBK --json               # → run.json content
+
+# Version check
+flow-walker --version                            # → flow-walker 0.1.0
+flow-walker --version --json                     # → {"version":"0.1.0"}
+```
+
+## Prerequisites
+
+1. **agent-flutter** installed and in PATH (`npm install -g agent-flutter-cli`)
+2. Flutter app running with Marionette initialized
+3. ADB connected (Android) or Simulator running (iOS)
+4. Run `agent-flutter doctor` to verify setup
+
+## Run IDs
+
+Every `flow-walker run` generates a unique **10-char base64url ID** (e.g. `25h7afGwBK`).
+
+- Output goes to `<output-dir>/<run-id>/` — multiple runs never overwrite
+- `run.json` includes `"id": "25h7afGwBK"` as top-level field
+- Agents can correlate runs by ID across logs, reports, and API calls
+- Composite key `{flow}/{id}` (e.g. `tab-navigation/25h7afGwBK`) for human reference
+
+## Canonical workflows
+
+### Auto-explore an app
+
+```bash
+agent-flutter connect ws://127.0.0.1:38047/abc=/ws
+flow-walker walk --skip-connect --max-depth 3 --output-dir ./flows/
+# Output: YAML flows + _nav-graph.json
+```
+
+### Execute a flow
+
+```bash
+flow-walker run flows/tab-navigation.yaml --output-dir ./results/
+# => Run ID: 25h7afGwBK
+# => Output: ./results/25h7afGwBK/run.json, recording.mp4, step-*.png, device.log
+```
+
+### Generate report
+
+```bash
+flow-walker report ./results/25h7afGwBK/
+# Output: report.html (self-contained, can be shared)
+```
+
+### Run a flow suite
+
+```bash
+for flow in flows/*.yaml; do
+  flow-walker run "$flow" --output-dir ./results/ --json
+done
+# Each run gets its own subdirectory by run ID
+```
+
+## Output shapes
+
+### run.json
+
+```json
+{
+  "id": "25h7afGwBK",
+  "flow": "tab-navigation",
+  "device": "Pixel_7a",
+  "startedAt": "2026-03-12T10:00:00Z",
+  "duration": 14200,
+  "result": "pass",
+  "steps": [
+    {
+      "index": 0,
+      "name": "Verify home tab",
+      "action": "assert",
+      "status": "pass",
+      "timestamp": 0,
+      "duration": 2300,
+      "elementCount": 22,
+      "screenshot": "step-1-tab-home.png",
+      "assertion": {
+        "interactive_count": { "min": 20, "actual": 22 },
+        "bottom_nav_tabs": { "min": 4, "actual": 4 }
+      }
+    }
+  ],
+  "video": "recording.mp4",
+  "log": "device.log"
+}
+```
+
+### _nav-graph.json (from walk)
+
+```json
+{
+  "nodes": [
+    { "id": "abc123", "name": "home-screen", "elementCount": 24, "visits": 3 }
+  ],
+  "edges": [
+    { "source": "abc123", "target": "def456", "element": { "ref": "@e3", "type": "button", "text": "Settings" } }
+  ]
+}
+```
+
+### Schema envelope (from schema)
+
+```json
+{
+  "version": "0.1.0",
+  "commands": [
+    {
+      "name": "run",
+      "description": "Execute a YAML flow...",
+      "args": [{ "name": "flow", "required": true, "type": "path", "description": "..." }],
+      "flags": [{ "name": "--json", "type": "boolean", "description": "..." }],
+      "exitCodes": { "0": "all steps pass", "1": "one or more steps fail", "2": "error" },
+      "examples": ["flow-walker run flows/tab-navigation.yaml"]
+    }
+  ]
+}
+```
+
+### push result (from push --json)
+
+```json
+{
+  "id": "25h7afGwBK",
+  "url": "https://flow-walker.beastoin.workers.dev/runs/25h7afGwBK",
+  "htmlUrl": "https://flow-walker.beastoin.workers.dev/runs/25h7afGwBK.html",
+  "expiresAt": "2026-04-11T13:22:12.070Z"
+}
+```
+
+### Command outputShape (from schema)
+
+Commands that produce structured output declare their fields via `outputShape`:
+
+```json
+{
+  "name": "run",
+  "outputShape": [
+    { "name": "id", "type": "string", "description": "Unique 10-char run ID" },
+    { "name": "flow", "type": "string", "description": "Flow name" },
+    { "name": "result", "type": "pass|fail", "description": "Overall result" },
+    { "name": "duration", "type": "number", "description": "Total milliseconds" },
+    { "name": "steps", "type": "StepResult[]", "description": "Per-step results" }
+  ]
+}
+```
+
+Commands with `outputShape`: `run`, `push`, `get`. Use `flow-walker schema <cmd>` to inspect.
+
+### Agent-readable run data
+
+After push, structured run data is available. URLs are agent-first — JSON by default:
+
+```bash
+# JSON (default) — for agents
+curl https://flow-walker.beastoin.workers.dev/runs/25h7afGwBK
+curl https://flow-walker.beastoin.workers.dev/runs/25h7afGwBK.json
+
+# HTML — for humans
+open https://flow-walker.beastoin.workers.dev/runs/25h7afGwBK.html
+```
+
+Returns run.json structure (without local file paths like video/screenshot filenames).
+
+CLI equivalent:
+
+```bash
+flow-walker get 25h7afGwBK          # pretty-printed JSON
+flow-walker get 25h7afGwBK --json   # compact JSON (pipe-friendly)
+flow-walker get 25h7afGwBK | jq '.steps[] | select(.status=="fail")'
+```
+
+### Structured error (on failure)
+
+```json
+{
+  "error": {
+    "code": "INVALID_INPUT",
+    "message": "Path contains traversal sequences",
+    "hint": "Remove .. from path",
+    "diagnosticId": "a1b2c3d4"
+  }
+}
+```
+
+## YAML flow format
+
+```yaml
+name: flow-name
+description: What this flow tests
+app: Omi                          # optional: app name
+app_url: https://omi.me           # optional: app URL
+covers:
+  - app/lib/pages/home.dart
+prerequisites:
+  - auth_ready
+setup: normal
+
+steps:
+  - name: Step description
+    press: { type: button, position: rightmost }
+    assert:
+      interactive_count: { min: 20 }
+      has_type: { type: switch, min: 2 }
+    screenshot: label
+```
+
+### Press targets
+
+| Target | Syntax |
+|--------|--------|
+| By ref | `{ ref: "@e3" }` |
+| By type | `{ type: button }` |
+| By position | `{ type: button, position: rightmost }` |
+| By nav tab | `{ bottom_nav_tab: 0 }` |
+
+### Assertions
+
+| Assertion | Syntax |
+|-----------|--------|
+| Element count | `interactive_count: { min: 20 }` |
+| Nav tabs | `bottom_nav_tabs: { min: 4 }` |
+| Element type | `has_type: { type: switch, min: 2 }` |
+| Text visible | `text_visible: ["Featured", "Home"]` |
+| Text absent | `text_not_visible: ["Error", "Sign In"]` |
+
+Text assertions use Android UIAutomator (via `agent-flutter text`) to check visible text from the accessibility layer. This captures text that Marionette snapshots miss (labels, content descriptions, system UI).
+
+## Exit codes
+
+| Code | Meaning |
+|------|---------|
+| `0` | Success |
+| `1` | Flow has failing steps |
+| `2` | Error (invalid args, file not found, device error) |
+
+## Error codes
+
+Every error returns `{"error": {"code": "...", "message": "...", "hint": "...", "diagnosticId": "..."}}`.
+
+| Code | Meaning | Common cause |
+|------|---------|-------------|
+| `INVALID_ARGS` | Bad CLI arguments | Missing required arg, unknown subcommand |
+| `INVALID_INPUT` | Input fails validation | Path traversal, control chars, bad URI format |
+| `FILE_NOT_FOUND` | Required file missing | No flow YAML, no run.json, remote run not found |
+| `FLOW_PARSE_ERROR` | Invalid YAML flow | Malformed YAML, missing name/steps |
+| `COMMAND_FAILED` | External command error | agent-flutter failure, network error, upload failure |
+
+## Environment variables
+
+Precedence: CLI flag > env var > default.
+
+| Variable | Purpose | Default |
+|----------|---------|---------|
+| `FLOW_WALKER_OUTPUT_DIR` | Default output directory | `./run-output/` |
+| `FLOW_WALKER_AGENT_PATH` | Path to agent-flutter binary | `agent-flutter` |
+| `FLOW_WALKER_DRY_RUN` | Enable dry-run mode | `0` |
+| `FLOW_WALKER_JSON` | Force JSON output | auto (TTY detection) |
+| `FLOW_WALKER_API_URL` | Hosted service URL for push/get | `https://flow-walker.beastoin.workers.dev` |
+| `AGENT_FLUTTER_DEVICE` | ADB device ID | auto-detect |
+
+JSON output precedence: `--no-json` > `--json` > `FLOW_WALKER_JSON=1` > TTY auto-detect (non-TTY defaults to JSON).

package/CLAUDE.md

@@ -0,0 +1,81 @@
+# CLAUDE.md — Working on flow-walker
+
+## Publish rule
+
+This repo is the **publish target** only. Source of truth is [`beastoin/autoloop`](https://github.com/beastoin/autoloop).
+All code changes must go through autoloop's phase-gated build loop first, then get copied here for npm publish.
+Do not edit this repo directly — add a new phase program in autoloop instead.
+
+## Project overview
+
+`flow-walker` is a Node.js CLI that auto-explores Flutter apps, executes YAML test flows, and generates HTML reports.
+It builds on [agent-flutter](https://github.com/beastoin/agent-flutter) for all device interaction.
+
+**Six commands:**
+- `walk` — BFS-explore the app, discover screens, generate YAML flows
+- `run` — Execute a YAML flow, produce run.json + video + screenshots
+- `report` — Generate self-contained HTML report from run results
+- `push` — Upload report to hosted service, return shareable URL
+- `get` — Fetch run data from hosted service by run ID
+- `schema` — Machine-readable command introspection
+
+**Design principles:**
+1. **agent-flutter as transport** — never touches VM Service or ADB directly
+2. **Fingerprint by structure** — screen identity uses element types/counts, not text
+3. **Safety first** — blocklist prevents pressing destructive elements
+4. **Self-contained output** — HTML reports embed everything as base64
+5. **YAML as contract** — flows are portable, readable, version-controllable
+
+## Architecture
+
+- `src/cli.ts` — entry point, arg parsing, subcommand routing
+- `src/walker.ts` — BFS exploration algorithm
+- `src/fingerprint.ts` — screen identity hashing
+- `src/graph.ts` — navigation graph data structure
+- `src/safety.ts` — blocklist evaluation
+- `src/yaml-writer.ts` — YAML flow generation
+- `src/agent-bridge.ts` — thin wrapper around agent-flutter CLI
+- `src/flow-parser.ts` — YAML → Flow object parsing
+- `src/runner.ts` — flow step execution engine
+- `src/reporter.ts` — HTML report generation
+- `src/capture.ts` — video, screenshot, logcat helpers
+- `src/run-schema.ts` — RunResult type + validation + run ID generation
+- `src/types.ts` — shared type definitions
+- `src/errors.ts` — structured error handling (FlowWalkerError)
+- `src/validate.ts` — input validation (paths, URIs, control chars)
+- `src/command-schema.ts` — command schema for agent discovery
+- `src/push.ts` — report upload to hosted service
+
+## Build and test
+
+```bash
+npm install
+npm test                    # all tests
+npx tsc --noEmit            # typecheck
+```
+
+## Code conventions
+
+- TypeScript ESM modules (`.ts` imports with explicit extension)
+- Node built-ins only; no external runtime dependencies
+- Node.js ≥ 22 with `--experimental-strip-types`
+- Exit codes: 0 = success, 1 = flow failure, 2 = error
+
+## Phase history
+
+| Phase | Focus | Eval |
+|-------|-------|------|
+| 1 | walk: BFS explorer, fingerprinting, safety, YAML generation | eval.sh (19 gates) |
+| 2 | run + report: flow executor, video/screenshots, HTML viewer | eval2.sh (25 gates) |
+| 3 | Agent-grade: structured errors, schema, input hardening, run IDs | eval3.sh (41 gates) |
+| 4 | Hosted reports: push command, Cloudflare Worker + R2 | eval4.sh (17 gates) |
+| 5 | Landing page: live metrics, stats tracking | eval5.sh (14 gates) |
+| 6 | Agent-friendly run data + app metadata | eval6.sh (13 gates) |
+
+## What not to do
+
+- Do not edit this repo directly — use autoloop
+- Do not add external runtime dependencies
+- Do not change exit code semantics
+- Do not bypass blocklist safety checks
+- Do not access VM Service or ADB directly (use agent-flutter)

package/LICENSE

@@ -0,0 +1,21 @@
+MIT License
+
+Copyright (c) 2026 beastoin
+
+Permission is hereby granted, free of charge, to any person obtaining a copy
+of this software and associated documentation files (the "Software"), to deal
+in the Software without restriction, including without limitation the rights
+to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+copies of the Software, and to permit persons to whom the Software is
+furnished to do so, subject to the following conditions:
+
+The above copyright notice and this permission notice shall be included in all
+copies or substantial portions of the Software.
+
+THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+SOFTWARE.

package/README.md

@@ -0,0 +1,247 @@
+# flow-walker
+
+Auto-discover app flows, execute YAML test flows, generate HTML reports.
+
+flow-walker is the **flow layer** — it defines, discovers, executes, and reports on flows. It uses [agent-flutter](https://github.com/beastoin/agent-flutter) and [agent-swift](https://github.com/beastoin/agent-swift) as **transport layers** that control specific platforms.
+
+```
+flow-walker (flows: walk, run, report, push, get, schema)
+    |
+agent-flutter (Flutter apps on Android/iOS)
+agent-swift   (native macOS/iOS apps)
+    |
+devices
+```
+
+## Quick start
+
+```bash
+npm install -g flow-walker-cli
+
+# Explore app automatically
+flow-walker walk --app-uri ws://127.0.0.1:38047/abc=/ws
+
+# Execute a specific flow
+flow-walker run flows/tab-navigation.yaml
+
+# Generate HTML report
+flow-walker report ./run-output/<run-id>/
+
+# Share report (hosted)
+flow-walker push ./run-output/<run-id>/
+
+# Retrieve run data
+flow-walker get 25h7afGwBK
+
+# Discover commands (agent-first)
+flow-walker schema
+flow-walker schema run
+
+# Version
+flow-walker --version
+```
+
+## Prerequisites
+
+- Node.js >= 22
+- [agent-flutter](https://github.com/beastoin/agent-flutter) installed and in PATH
+- Flutter app running with Marionette initialized
+- ADB connected (Android) or Simulator running (iOS)
+
+## Commands
+
+### `walk` — Auto-explore
+
+Discovers screens by pressing every interactive element, building a navigation graph, and generating YAML flow files.
+
+```bash
+flow-walker walk --app-uri ws://... --max-depth 3 --output-dir ./flows/
+flow-walker walk --skip-connect --json    # NDJSON: one event per line
+flow-walker walk --dry-run                # plan without pressing
+```
+
+### `run` — Execute flow
+
+Runs a YAML flow step-by-step. Each run gets a **unique ID** (10-char base64url like `P-tnB_sgKA`). Output goes to `<output-dir>/<run-id>/` so multiple runs never overwrite each other.
+
+```bash
+flow-walker run flows/tab-navigation.yaml
+# => Run ID: 25h7afGwBK
+# => Output: ./run-output/25h7afGwBK/
+
+flow-walker run flows/login.yaml --json   # machine-readable
+flow-walker run flows/settings.yaml --dry-run  # parse + resolve without executing
+```
+
+Output per run:
+- `run.json` — structured results with run ID, per-step status, timing, assertions
+- `recording.mp4` — screen recording with step timestamps
+- `step-N-*.png` — per-step screenshots
+- `device.log` — filtered device logs
+
+### `report` — Generate HTML viewer
+
+Self-contained HTML with embedded video, screenshots, and clickable step timeline.
+
+```bash
+flow-walker report ./run-output/25h7afGwBK/
+```
+
+### `push` — Share report
+
+Uploads report.html to the hosted service and returns a shareable URL. No auth, no config.
+
+```bash
+flow-walker push ./run-output/25h7afGwBK/
+# => URL: https://flow-walker.beastoin.workers.dev/runs/25h7afGwBK
+
+flow-walker push ./run-output/25h7afGwBK/ --json
+# => {"id":"25h7afGwBK","url":"https://...","htmlUrl":"https://....html","expiresAt":"2026-04-11T..."}
+```
+
+Reports are stored for 30 days. Re-pushing the same run is idempotent — returns the same URL with updated expiry. Use `FLOW_WALKER_API_URL` env var to point at a custom server.
+
+Push also uploads `run.json` (stripped of local file paths). URL scheme is agent-first:
+
+```bash
+# Agent: JSON by default
+curl https://flow-walker.beastoin.workers.dev/runs/25h7afGwBK
+curl https://flow-walker.beastoin.workers.dev/runs/25h7afGwBK.json
+
+# Human: HTML report
+open https://flow-walker.beastoin.workers.dev/runs/25h7afGwBK.html
+```
+
+### `get` — Retrieve run data
+
+Fetches structured run data from the hosted service by run ID. Returns JSON (the same run.json uploaded during push, minus local file paths).
+
+```bash
+flow-walker get 25h7afGwBK          # pretty-printed
+flow-walker get 25h7afGwBK --json   # compact (pipe-friendly)
+flow-walker get 25h7afGwBK | jq '.steps[] | select(.status=="fail")'
+```
+
+### `schema` — Agent discovery
+
+Machine-readable command introspection. Returns versioned JSON with args, flags (with types), exit codes, and examples.
+
+```bash
+flow-walker schema           # all commands with version envelope
+flow-walker schema run       # single command detail
+```
+
+Agents can discover capabilities programmatically — no --help parsing needed.
+
+## YAML flow format
+
+```yaml
+name: tab-navigation
+description: Bottom nav bar detection, switch between 4 tabs
+app: Omi                                    # optional: app name (shown in reports)
+app_url: https://omi.me                     # optional: app URL (linked in reports)
+covers:
+  - app/lib/pages/home/page.dart
+prerequisites:
+  - auth_ready
+setup: normal
+
+steps:
+  - name: Verify home tab and nav bar
+    assert:
+      interactive_count: { min: 20 }
+      bottom_nav_tabs: { min: 4 }
+    screenshot: tab-home
+
+  - name: Switch to tab 2
+    press: { bottom_nav_tab: 1 }
+    screenshot: tab-2
+
+  - name: Scroll in current tab
+    scroll: down
+
+  - name: Verify developer settings has switches
+    assert:
+      has_type: { type: switch, min: 2 }
+
+  - name: Verify visible text on screen
+    assert:
+      text_visible: ["Featured", "Home"]
+      text_not_visible: ["Error", "Sign In"]
+
+  - name: Return to home tab
+    press: { bottom_nav_tab: 0 }
+    screenshot: final
+```
+
+### Step actions
+
+| Action | Syntax | Description |
+|--------|--------|-------------|
+| `press` | `{ type: button, position: rightmost }` | Press by type, position, ref, or bottom_nav_tab |
+| `scroll` | `down` / `up` / `left` / `right` | Scroll the screen |
+| `fill` | `{ type: textfield, value: "text" }` | Enter text into a field |
+| `back` | `true` | Navigate back |
+| `assert` | `{ interactive_count: { min: N } }` | Assert element counts, nav tabs, element types |
+| `screenshot` | `name` | Capture screenshot with label |
+
+### Assertions
+
+| Assertion | Syntax | Description |
+|-----------|--------|-------------|
+| `interactive_count` | `{ min: 20 }` | Min total interactive elements on screen |
+| `bottom_nav_tabs` | `{ min: 4 }` | Min bottom navigation tabs |
+| `has_type` | `{ type: switch, min: 2 }` | Min elements of a specific type |
+| `text_visible` | `["Featured", "Home"]` | Text must be visible on screen (via UIAutomator) |
+| `text_not_visible` | `["Error", "Sign In"]` | Text must NOT be visible on screen |
+
+## Agent-friendly design
+
+Built following [Poehnelt's CLI-for-agents principles](https://justin.poehnelt.com/posts/rewrite-your-cli-for-ai-agents/):
+
+- **Schema introspection** — `flow-walker schema` returns versioned JSON with typed args/flags
+- **Structured errors** — every error returns `{code, message, hint, diagnosticId}` in JSON
+- **Input hardening** — path traversal, control chars, URI format all validated
+- **TTY-aware JSON** — `--no-json` > `--json` > `FLOW_WALKER_JSON=1` > TTY auto-detect
+- **Dry-run** — `--dry-run` parses and resolves targets without executing (includes resolve reasons)
+- **NDJSON streaming** — walk emits `walk:start`, `screen`, `edge`, `skip` events as one JSON per line
+- **Unique run IDs** — 10-char base64url per run, filesystem-safe, URL-safe
+- **Hosted sharing** — `flow-walker push` uploads report and returns a URL, no auth needed
+- **Agent-first URLs** — `/runs/:id` defaults to JSON; `.html` suffix for humans
+- **App metadata** — optional `app` + `app_url` in YAML flows, shown in reports and landing page
+- **Environment variables** — `FLOW_WALKER_OUTPUT_DIR`, `FLOW_WALKER_AGENT_PATH`, `FLOW_WALKER_DRY_RUN`, `FLOW_WALKER_JSON`, `FLOW_WALKER_API_URL`
+- **Exit codes** — 0 = success, 1 = flow failure, 2 = error
+
+## Architecture
+
+```
+flow-walker CLI (flow layer)
+  | shells out to
+agent-flutter CLI / agent-swift CLI (transport layer)
+  | connects via
+VM Service + Marionette / XCTest (platform-specific)
+  | controls
+App on device/emulator/desktop
+```
+
+**Design principles:**
+1. **Pluggable transport** — same YAML flows, different backends
+2. **Fingerprint by structure** — screen identity uses element types/counts, not text
+3. **Safety first** — blocklist prevents pressing destructive elements
+4. **Self-contained output** — HTML reports embed everything as base64
+5. **YAML as contract** — flows are portable, readable, version-controllable
+
+## Phase history
+
+| Phase | Focus | Status |
+|-------|-------|--------|
+| 1 | walk: BFS explorer, fingerprinting, safety, YAML generation | Complete |
+| 2 | run + report: flow executor, video/screenshots, HTML viewer | Complete |
+| 3 | Agent-grade: structured errors, schema, input hardening, run IDs | Complete |
+| 4 | Hosted reports: push command, Cloudflare Worker + R2 | Complete |
+| 5 | Landing page: live metrics, stats tracking | Complete |
+| 6 | Agent-friendly run data + app metadata | Complete |
+
+## License
+
+MIT

package/package.json

@@ -0,0 +1,21 @@
+{
+  "name": "flow-walker-cli",
+  "version": "0.2.0",
+  "description": "Automatic app flow extraction via agent-flutter",
+  "type": "module",
+  "bin": {
+    "flow-walker": "./src/cli.ts"
+  },
+  "scripts": {
+    "test": "node --experimental-strip-types --test tests/*.test.ts",
+    "build": "tsc --noEmit"
+  },
+  "engines": {
+    "node": ">=22"
+  },
+  "license": "MIT",
+  "devDependencies": {
+    "typescript": "^5.7.0",
+    "@types/node": "^22.0.0"
+  }
+}