npm - monomind - Versions diffs - 1.19.0 → 2.0.1 - Mend

monomind 1.19.0 → 2.0.1

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.

Files changed (883) hide show

package/.claude/skills/monomind/browse.md DELETED Viewed

@@ -1,724 +0,0 @@
----
-name: monomind:browse
-description: State-of-the-art browser automation skill for UI testing, web scraping, and agent-driven navigation using monomind browse (built-in TypeScript CDP client). Token-optimized with ref-based element selection, batch execution, KV-cache prompt ordering, on-demand screenshots, and full monomind memory integration.
-version: 2.0.0
-triggers:
-  - /browse
-  - monomind:browse
-  - browse the web
-  - test the UI
-  - browser automation
-  - open the browser
-  - navigate to
-  - click on the website
-  - fill out the form
-  - test login
-  - take a screenshot
-  - check the page
-  - web vitals
-  - react tree
-  - scrape
-  - crawl
-tools:
-  - Bash
-requires:
-  - monomind >= 1.0.0
----
-# monomind:browse
-State-of-the-art browser automation using monomind browse. Optimized for minimal token consumption, maximum test coverage, and deep monomind integration.
----
-## Setup (Run Once)
-```bash
-# Install
-# monomind browse is built-in — no install needed
-# Download Chrome (first time only)
-# no install needed
-# Verify
-npx monomind browse --version   # should be >= 0.25.4
-npx monomind browse doctor      # check all systems
-```
----
-## Token Efficiency Rules (ALWAYS FOLLOW)
-These rules are non-negotiable. Violating them wastes tokens and degrades performance.
-1. **Batch multi-step flows** — Use `monomind browse batch` to execute sequences in a single process invocation. Eliminates per-command startup overhead.
-2. **Snapshot with `-i` flag** — Interactive-only snapshots are 93% smaller than full trees. Never call `monomind browse snapshot` without `-i` unless you need to understand full page structure.
-3. **Reuse refs** — After a snapshot, refs (`@e1`, `@e2`) are stable for the current page. Do NOT re-snapshot unless the page changed. One snapshot per page-state.
-4. **On-demand screenshots only** — Screenshots add ~800ms and ~1500 tokens as images. Only call `monomind browse screenshot` when: element not in a11y tree, visual verification is required, or the task explicitly needs visual proof.
-5. **Use batch for read-only chains** — `npx monomind browse open url` then `snapshot -i` is two process starts. `monomind browse batch "open url" "snapshot -i"` is one.
-6. **Scope snapshots** — When testing a form or component, use `monomind browse snapshot -i "#form-id"` to scope to that subtree only.
-7. **Prefer `wait --text` over polling** — Never sleep and re-snapshot. Use `monomind browse wait --text "Expected"` or `wait --url "**pattern"`.
----
-## Core Loop
-```
-OPEN → SNAPSHOT -i → ACT (by ref) → SNAPSHOT -i (if page changed) → VERIFY → REPEAT
-```
-### Minimal golden-path example
-```bash
-# Batch: open + snapshot in one call
-npx monomind browse batch "open https://app.example.com/login" "snapshot -i"
-# → snapshot output: textbox "Email" [ref=e1], textbox "Password" [ref=e2], button "Sign In" [ref=e3]
-npx monomind browse fill @e1 "user@test.com"
-npx monomind browse fill @e2 "SecurePass123!"
-npx monomind browse click @e3
-npx monomind browse wait --url "**/dashboard" --timeout 8000
-# Re-snapshot only because URL changed (new page state)
-npx monomind browse snapshot -i
-```
-### Full batch mode (most token-efficient)
-```bash
-npx monomind browse batch \
-  "open https://app.example.com/login" \
-  "snapshot -i" \
-  "fill @e1 user@test.com" \
-  "fill @e2 SecurePass123!" \
-  "click @e3" \
-  "wait --url **/dashboard" \
-  "snapshot -i"
-```
-Or as JSON via stdin (best for programmatic use):
-```bash
-echo '[
-  ["open", "https://app.example.com/login"],
-  ["snapshot", "-i"],
-  ["fill", "@e1", "user@test.com"],
-  ["fill", "@e2", "SecurePass123!"],
-  ["click", "@e3"],
-  ["wait", "--url", "**/dashboard"],
-  ["snapshot", "-i"]
-]' | monomind browse batch --json --bail
-```
----
-## All Commands Reference
-### Navigation
-```bash
-npx monomind browse open <url>              # Launch + navigate (aliases: goto, navigate)
-npx monomind browse open                    # Launch on about:blank (for pre-nav setup)
-npx monomind browse back                    # Go back
-npx monomind browse forward                 # Go forward
-npx monomind browse reload                  # Reload
-npx monomind browse pushstate <url>         # SPA client-side nav (Next.js, Remix, etc.)
-npx monomind browse close                   # Close browser
-npx monomind browse close --all             # Close all sessions
-```
-### Snapshot (primary observation tool)
-```bash
-npx monomind browse snapshot                # Full accessibility tree + refs
-npx monomind browse snapshot -i             # Interactive elements only (USE THIS BY DEFAULT)
-npx monomind browse snapshot -i --urls      # Include href URLs for links
-npx monomind browse snapshot -c             # Compact (strip empty structural nodes)
-npx monomind browse snapshot -d 3           # Limit depth to 3 levels
-npx monomind browse snapshot -s "#main"     # Scope to CSS selector
-npx monomind browse snapshot -i -c -d 5    # Combined: compact interactive, max depth 5
-npx monomind browse snapshot --json         # JSON output for programmatic use
-```
-**Output example:**
-```
-- textbox "Email" [ref=e1]
-- textbox "Password" [ref=e2]
-- button "Sign In" [ref=e3]
-- link "Forgot password" [ref=e4]
-```
-### Interaction
-```bash
-# Primary: use @ref from snapshot (fastest, no DOM re-query)
-npx monomind browse click @e3
-npx monomind browse fill @e1 "value"
-npx monomind browse dblclick @e5
-npx monomind browse hover @e6
-npx monomind browse check @e7               # checkbox
-npx monomind browse uncheck @e8
-npx monomind browse select @e9 "Option A"
-npx monomind browse focus @e2
-# Keyboard
-npx monomind browse press Enter
-npx monomind browse press Tab
-npx monomind browse press "Control+a"
-npx monomind browse keyboard type "hello world"      # real keystrokes
-npx monomind browse keyboard inserttext "hello"      # insert without key events
-npx monomind browse keydown Shift
-npx monomind browse keyup Shift
-# Drag & drop
-npx monomind browse drag @e1 @e2
-npx monomind browse upload @e3 /path/to/file.pdf
-# Scroll
-npx monomind browse scroll down 500         # scroll 500px down
-npx monomind browse scroll up
-npx monomind browse scrollintoview @e10
-npx monomind browse scroll down --selector "#feed"
-```
-### Semantic locators (fallback when no ref available)
-```bash
-npx monomind browse find role button click --name "Submit"
-npx monomind browse find text "Sign In" click
-npx monomind browse find label "Email" fill "test@test.com"
-npx monomind browse find placeholder "Search..." fill "query"
-npx monomind browse find testid "submit-btn" click
-npx monomind browse find first ".item" click
-npx monomind browse find nth 2 "a" text
-```
-### Wait (never sleep/poll manually)
-```bash
-npx monomind browse wait 500                         # ms delay (use sparingly)
-npx monomind browse wait "#spinner" --state hidden   # wait for element to hide
-npx monomind browse wait --text "Success"            # wait for text to appear
-npx monomind browse wait --url "**/dashboard"        # wait for URL pattern
-npx monomind browse wait --load networkidle          # wait for network idle
-npx monomind browse wait --fn "window.ready === true"  # wait for JS condition
-npx monomind browse wait --fn "!document.body.innerText.includes('Loading')"
-```
-### Read state
-```bash
-npx monomind browse get text @e1           # text content
-npx monomind browse get html @e2           # innerHTML
-npx monomind browse get value @e3          # input value
-npx monomind browse get attr @e4 "href"    # attribute
-npx monomind browse get title              # page title
-npx monomind browse get url                # current URL
-npx monomind browse get count ".item"      # count matching elements
-npx monomind browse get box @e1            # bounding box
-npx monomind browse get styles @e1         # computed CSS styles
-npx monomind browse is visible @e1         # boolean check
-npx monomind browse is enabled @e1
-npx monomind browse is checked @e1
-```
-### Screenshots (use sparingly)
-```bash
-npx monomind browse screenshot                        # auto-path in /tmp
-npx monomind browse screenshot page.png              # specific path
-npx monomind browse screenshot --full full-page.png  # full-page scroll capture
-npx monomind browse screenshot --annotate            # numbered refs overlaid → use with visual models
-npx monomind browse pdf report.pdf                   # save as PDF
-```
-**Annotated screenshot pattern** (for visual debugging or multimodal LLMs):
-```bash
-npx monomind browse screenshot --annotate
-# Output: [1] @e1 button "Submit"  [2] @e2 link "Home"  [3] @e3 textbox "Email"
-# Now refs are cached — interact immediately without re-snapshot
-npx monomind browse click @e1
-```
-### Diff (regression testing)
-```bash
-npx monomind browse diff snapshot                           # current vs last snapshot
-npx monomind browse diff snapshot --baseline ./before.txt   # vs saved file
-npx monomind browse diff snapshot -s "#main" --compact      # scoped diff
-npx monomind browse diff screenshot --baseline before.png   # pixel diff
-npx monomind browse diff screenshot --baseline b.png -t 0.2 # threshold 0–1
-npx monomind browse diff url https://v1.com https://v2.com  # compare two URLs
-npx monomind browse diff url https://v1.com https://v2.com --screenshot  # + visual
-```
-### Tabs & multi-tab
-```bash
-npx monomind browse tab                          # list all tabs
-npx monomind browse tab new https://example.com  # new tab
-npx monomind browse tab new --label docs https://docs.example.com  # named tab
-npx monomind browse tab docs                     # switch by label
-npx monomind browse tab t2                       # switch by stable id
-npx monomind browse tab close docs               # close by label
-npx monomind browse window new                   # new window
-npx monomind browse frame "#iframe-id"           # switch to iframe
-npx monomind browse frame main                   # back to main frame
-```
-**Multi-tab parallel test pattern:**
-```bash
-npx monomind browse tab new --label app https://app.example.com
-npx monomind browse tab new --label docs https://docs.example.com
-npx monomind browse tab app
-npx monomind browse snapshot -i    # refs for app tab
-npx monomind browse click @e3
-npx monomind browse tab docs
-npx monomind browse snapshot -i    # refs for docs tab
-```
-### Dialogs
-```bash
-npx monomind browse dialog accept "confirmation text"
-npx monomind browse dialog dismiss
-npx monomind browse dialog status            # is a dialog currently open?
-```
-Note: `alert` and `beforeunload` are auto-accepted by default. `confirm` and `prompt` need explicit handling.
-### Network interception
-```bash
-npx monomind browse network route "https://api.example.com/*" --abort    # block endpoint
-npx monomind browse network route "*" --abort --resource-type script      # block all JS
-npx monomind browse network route "https://api/*" --body '{"data":[]}'   # mock response
-npx monomind browse network unroute "https://api.example.com/*"
-npx monomind browse network requests                          # view tracked requests
-npx monomind browse network requests --filter api             # filter by URL substring
-npx monomind browse network requests --type xhr,fetch         # filter by type
-npx monomind browse network requests --method POST
-npx monomind browse network requests --status 2xx
-npx monomind browse network request <requestId>               # full request/response
-npx monomind browse network har start                         # record HAR
-npx monomind browse network har stop output.har               # stop + save
-```
-**Pre-nav setup pattern** (set network routes BEFORE navigating):
-```bash
-npx monomind browse batch \
-  '["open"]' \
-  '["network", "route", "*", "--abort", "--resource-type", "script"]' \
-  '["cookies", "set", "--curl", "auth.curl", "--domain", "localhost"]' \
-  '["navigate", "http://localhost:3000"]'
-```
-### Cookies & storage
-```bash
-npx monomind browse cookies                              # get all cookies
-npx monomind browse cookies set name value              # set cookie
-npx monomind browse cookies set --curl cookies.curl     # import from cURL dump / JSON / header string
-npx monomind browse cookies clear
-npx monomind browse storage local                        # get localStorage
-npx monomind browse storage local myKey                  # get specific key
-npx monomind browse storage local set myKey myValue      # set
-npx monomind browse storage local clear
-npx monomind browse storage session                      # sessionStorage (same API)
-```
-### Browser settings
-```bash
-npx monomind browse set viewport 1280 720 2             # width height deviceScaleFactor
-npx monomind browse set device "iPhone 15 Pro"          # device emulation
-npx monomind browse set geo 37.7749 -122.4194           # geolocation
-npx monomind browse set offline on                      # offline mode
-npx monomind browse set headers '{"Authorization":"Bearer tok"}'  # global headers
-npx monomind browse set credentials user pass           # HTTP basic auth
-npx monomind browse set media dark                      # color scheme
-```
-### Clipboard
-```bash
-npx monomind browse clipboard read
-npx monomind browse clipboard write "Hello"
-npx monomind browse clipboard copy      # Ctrl+C
-npx monomind browse clipboard paste     # Ctrl+V
-```
-### Mouse (raw control, use only when refs/semantic locators fail)
-```bash
-npx monomind browse mouse move 100 200
-npx monomind browse mouse down left
-npx monomind browse mouse up left
-npx monomind browse mouse wheel 100 0   # dy dx
-```
-### React DevTools (v0.27+)
-Requires launching with `--enable react-devtools`:
-```bash
-# Launch with hook installed
-npx monomind browse open --enable react-devtools https://your-react-app.com
-# Inspect component tree
-npx monomind browse react tree                         # full component hierarchy
-npx monomind browse react inspect 5                    # fiber ID → props, hooks, state, source
-npx monomind browse react renders start                # begin render profiling
-npx monomind browse react renders stop                 # print profile (mount/re-render counts)
-npx monomind browse react renders stop --json          # JSON output
-npx monomind browse react suspense                     # Suspense boundaries + root-cause classifier
-npx monomind browse react suspense --only-dynamic      # hide static boundaries
-```
-### Web Vitals (framework-agnostic)
-```bash
-npx monomind browse vitals                             # LCP, CLS, TTFB, FCP, INP + React hydration
-npx monomind browse vitals https://example.com         # test specific URL
-npx monomind browse vitals --json                      # JSON output
-```
-### Tracing & profiling
-```bash
-npx monomind browse trace start trace.zip             # start Chrome trace
-npx monomind browse trace stop trace.zip              # stop and save
-npx monomind browse profiler start                    # DevTools profiler
-npx monomind browse profiler stop profile.json        # stop and save
-npx monomind browse console                           # browser console messages
-npx monomind browse console --json                    # structured CDP output
-npx monomind browse console --clear
-npx monomind browse errors                            # uncaught JS exceptions
-npx monomind browse errors --clear
-```
-### Sessions & auth
-```bash
-# Isolated sessions (each has own browser, cookies, history)
-npx monomind browse --session agent1 open site-a.com
-npx monomind browse --session agent2 open site-b.com
-npx monomind browse session list
-# Persist state across restarts
-npx monomind browse --session-name myapp open app.example.com
-# → auto-saves to ~/.monomind/browser-sessions/myapp
-# Reuse existing Chrome login
-npx monomind browse profiles                          # list Chrome profiles
-npx monomind browse --profile Default open gmail.com
-# Save / load state
-npx monomind browse state save ./auth.json            # save cookies + localStorage
-npx monomind browse state load ./auth.json
-npx monomind browse --state ./auth.json open https://app.example.com/dashboard
-# Auth vault (credentials never sent to LLM)
-echo "mypassword" | monomind browse auth save github --url https://github.com/login --username me --password-stdin
-npx monomind browse auth login github
-# Encrypted state at rest
-export AGENT_BROWSER_ENCRYPTION_KEY=<64-char-hex>
-npx monomind browse --session-name secure open example.com
-```
-### Dashboard (observability)
-```bash
-npx monomind browse dashboard start              # port 4848
-npx monomind browse dashboard start --port 8080
-npx monomind browse dashboard stop
-# → open http://localhost:4848 for live viewport + activity feed + AI chat
-```
-### Init scripts
-```bash
-npx monomind browse open --init-script ./setup.js https://app.example.com
-npx monomind browse addinitscript "window.__TEST__ = true"
-npx monomind browse removeinitscript <identifier>
-```
-### iOS / Mobile (real Safari)
-Requires: `npm install -g appium && appium driver install xcuitest`
-```bash
-npx monomind browse device list
-npx monomind browse -p ios --device "iPhone 15 Pro" open https://example.com
-npx monomind browse -p ios snapshot -i
-npx monomind browse -p ios tap @e1
-npx monomind browse -p ios fill @e2 "text"
-npx monomind browse -p ios swipe up
-npx monomind browse -p ios screenshot mobile.png
-npx monomind browse -p ios close
-```
-### Cloud providers
-| Provider | Env var | Flag |
-|---------|---------|------|
-| Browserbase | `BROWSERBASE_API_KEY` | `-p browserbase` |
-| Browser Use | `BROWSER_USE_API_KEY` | `-p browseruse` |
-| Browserless | `BROWSERLESS_API_KEY` | `-p browserless` |
-| Kernel | `KERNEL_API_KEY` | `-p kernel` |
-| AWS AgentCore | AWS credentials | `-p agentcore` |
-All commands work identically regardless of provider.
-### Streaming
-```bash
-npx monomind browse stream status           # see WebSocket port
-npx monomind browse stream enable --port 9223
-npx monomind browse stream disable
-```
-### CDP / Electron apps
-```bash
-npx monomind browse connect 9222            # connect to port, persist for session
-npx monomind browse --cdp 9222 snapshot     # per-command
-npx monomind browse --cdp wss://remote/cdp snapshot
-npx monomind browse --auto-connect snapshot  # auto-discover running Chrome
-```
----
-## Test Flows
-### Login / Auth
-```bash
-npx monomind browse batch \
-  "open https://app.example.com/login" \
-  "snapshot -i"
-# identify refs from output, then:
-npx monomind browse fill @e[email] "user@test.com"
-npx monomind browse fill @e[password] "TestPass123!"
-npx monomind browse click @e[submit]
-npx monomind browse wait --url "**/dashboard" --timeout 8000
-```
-### Form with validation
-```bash
-# Test happy path
-npx monomind browse batch "open /form" "snapshot -i"
-npx monomind browse fill @e1 "John Doe"
-npx monomind browse fill @e2 "john@test.com"
-npx monomind browse select @e3 "Option A"
-npx monomind browse check @e4
-npx monomind browse click @e5
-npx monomind browse wait --text "submitted"
-npx monomind browse screenshot pass-form.png
-# Test validation (empty submit)
-npx monomind browse reload
-npx monomind browse snapshot -i
-npx monomind browse click @e5                    # submit empty
-npx monomind browse wait --text "required"
-npx monomind browse snapshot -i                  # verify error messages
-# Invalid email
-npx monomind browse fill @e2 "not-an-email"
-npx monomind browse click @e5
-npx monomind browse snapshot -i
-```
-### CRUD
-```bash
-# Create
-npx monomind browse click @e[add]
-npx monomind browse fill @e[name] "New Item"
-npx monomind browse click @e[save]
-npx monomind browse wait --text "New Item"
-# Update
-npx monomind browse click @e[edit]
-npx monomind browse fill @e[name] "Updated Item"
-npx monomind browse click @e[save]
-npx monomind browse wait --text "Updated Item"
-# Delete
-npx monomind browse click @e[delete]
-npx monomind browse wait --text "Are you sure"
-npx monomind browse click @e[confirm]
-npx monomind browse wait --fn "!document.body.innerText.includes('Updated Item')"
-```
-### Multi-step wizard
-```bash
-# Step 1
-npx monomind browse batch "open /wizard" "snapshot -i"
-npx monomind browse fill @e1 "value"
-npx monomind browse click @e[next]
-npx monomind browse wait --text "Step 2"
-# Step 2
-npx monomind browse snapshot -i
-npx monomind browse select @e2 "choice"
-npx monomind browse click @e[next]
-# Step 3 — verify summary
-npx monomind browse snapshot -i
-npx monomind browse get text @e[summary]
-npx monomind browse click @e[confirm]
-npx monomind browse wait --text "Complete"
-```
-### Regression test (diff baseline)
-```bash
-# Save baseline
-npx monomind browse open https://app.example.com
-npx monomind browse snapshot -i > baseline.txt
-# After a deploy, compare:
-npx monomind browse open https://app.example.com
-npx monomind browse diff snapshot --baseline ./baseline.txt
-```
-### API mocking
-```bash
-# Mock API response, test UI reaction
-npx monomind browse batch \
-  '["open"]' \
-  '["network", "route", "https://api.example.com/users", "--body", "{\"data\":[]}"]' \
-  '["navigate", "https://app.example.com/users"]'
-npx monomind browse snapshot -i
-# → verify "No users found" empty state renders correctly
-```
-### React app deep inspection
-```bash
-npx monomind browse open --enable react-devtools https://your-react-app.com
-npx monomind browse react tree
-npx monomind browse vitals --json
-npx monomind browse react renders start
-# ... trigger user interactions ...
-npx monomind browse react renders stop
-npx monomind browse react suspense --only-dynamic
-```
----
-## Configuration
-Create in project root for persistent defaults:
-```json
-{
-  "$schema": "https://monomind.dev/schema.json",
-  "maxOutput": 50000,
-  "contentBoundaries": true,
-  "idleTimeout": "5m",
-  "screenshotDir": "./screenshots",
-  "screenshotFormat": "jpeg",
-  "screenshotQuality": 80
-}
-```
-Key security defaults for agent deployments:
-```json
-{
-  "contentBoundaries": true,
-  "maxOutput": 50000,
-  "allowedDomains": ["app.example.com", "*.example.com"],
-  "noAutoDialog": false
-}
-```
-Key env vars:
-```bash
-AGENT_BROWSER_SESSION=<name>          # session isolation
-AGENT_BROWSER_SESSION_NAME=<name>     # auto-persist state
-AGENT_BROWSER_MAX_OUTPUT=50000        # prevent context flooding
-AGENT_BROWSER_DEFAULT_TIMEOUT=30000  # op timeout in ms (default: 25000)
-AGENT_BROWSER_IDLE_TIMEOUT_MS=300000  # daemon auto-shutdown after idle
-AGENT_BROWSER_CONTENT_BOUNDARIES=1   # LLM-safe output delimiters
-AGENT_BROWSER_HEADED=1               # visible browser (debugging)
-AGENT_BROWSER_STREAM_PORT=9223       # fixed WebSocket stream port
-# No API key needed — monomind browse is built-in
-AI_GATEWAY_MODEL=anthropic/claude-sonnet-4-6
-```
----
-## Monomind Integration
-### Store successful test flows
-```bash
-npx monomind memory store \
-  --namespace browse \
-  --key "login-flow-<app>" \
-  --value "open /login → snapshot -i → fill @e[email] → fill @e[pw] → click @e[submit] → wait **/dashboard"
-```
-### Retrieve before testing
-```bash
-npx monomind memory search --query "login flow" --namespace browse
-```
-### Report bugs as tasks
-```bash
-npx monomind task create \
-  --title "UI Bug: form submits with empty email" \
-  --description "Steps: open /login, click submit without filling email. No validation shown. Screenshot: /tmp/bug-123.png"
-```
-### Save auth state for reuse across sessions
-```bash
-# Once logged in:
-npx monomind browse state save .monomind/auth/<app>.json
-# Future sessions:
-npx monomind browse --state .monomind/auth/<app>.json open https://app.example.com
-```
----
-## Anti-patterns (NEVER DO)
-| Anti-pattern | Why | Fix |
-|---|---|---|
-| `monomind browse snapshot` (no `-i`) for every step | Full tree = 10–20x tokens | Use `snapshot -i` always |
-| Playwright MCP | 13,700-token schema tax before step 1 | Use monomind browse directly |
-| Screenshot every step | +800ms +1500 tokens each | Screenshot only on fail/visual-required |
-| Re-snapshot without page change | Wastes tokens | Reuse refs from last snapshot |
-| `sleep N` between actions | Slow, fragile | Use `wait --text`, `wait --url`, `wait --fn` |
-| CSS selectors when refs available | Slower, can break | Always prefer `@eN` refs from snapshot |
-| Separate commands when batch works | Extra process starts | Use `batch` for multi-step flows |
----
-## Checklist
-When this skill is activated:
-- [ ] `npx monomind browse --version` — confirm monomind is installed
-- [ ] `npx monomind doctor` — check system health
-- [ ] Get target URL from user if not provided
-- [ ] Use `batch "open <url>" "snapshot -i"` to start
-- [ ] Use refs (`@eN`) from snapshot output for all interactions
-- [ ] Only re-snapshot after confirmed page-state change
-- [ ] Only screenshot when visual evidence is required
-- [ ] Use `wait --text / --url / --fn` instead of sleep or polling
-- [ ] Report results: ✓ PASS / ✗ FAIL (steps to reproduce) / ⚠ WARN
-- [ ] Store successful patterns in monomind memory (`browse` namespace)
-- [ ] Create monomind task for any found bugs