@aarwitz/tapp 0.17.0-rc.8 → 0.17.0
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/.claude-plugin/marketplace.json +34 -0
- package/.claude-plugin/plugin.json +33 -0
- package/AGENTS.md +34 -14
- package/Harness/OCQAHarnessUITests/ExplorerTests.swift +11 -0
- package/README.md +126 -77
- package/bin/tapp.js +187 -63
- package/docs/BROWSER-PRODUCT.md +1 -1
- package/docs/application-model.md +12 -3
- package/mcp-server/src/android-explorer.js +3 -1
- package/mcp-server/src/android-flow.js +18 -1
- package/mcp-server/src/application-model.js +83 -13
- package/mcp-server/src/ci-report.js +2 -2
- package/mcp-server/src/ci-setup.js +4 -4
- package/mcp-server/src/environment-preflight.js +43 -0
- package/mcp-server/src/html-report.js +3 -2
- package/mcp-server/src/index.js +326 -110
- package/mcp-server/src/pr-selection.js +2 -2
- package/mcp-server/src/product-operations.js +111 -6
- package/mcp-server/src/report.js +17 -4
- package/mcp-server/src/web-explorer.js +123 -10
- package/mcp-server/src/web-flow.js +17 -1
- package/package.json +6 -4
- package/scripts/ci-gate.sh +42 -0
- package/scripts/flow_lib.py +1 -1
- package/scripts/quick-capture.sh +2 -2
- package/skills/tapp/SKILL.md +75 -0
- package/skills/tapp/agents/openai.yaml +4 -0
- package/skills/tapp/references/commands.md +102 -0
|
@@ -0,0 +1,75 @@
|
|
|
1
|
+
---
|
|
2
|
+
name: tapp
|
|
3
|
+
description: Use Tapp to see, drive, explore, and verify real application surfaces on iOS simulators, Android emulators/devices, or the web. Use when a user asks an agent to test an app or UI change, find bugs, inspect or screenshot a screen, exercise a journey, create a replayable flow, gather release evidence, or run the deterministic Tapp gate. Also use when the user mentions Tapp, @aarwitz/tapp, tapp_* tools, .tapp artifacts, or asks whether agent-authored UI actually works.
|
|
4
|
+
---
|
|
5
|
+
|
|
6
|
+
# Tapp
|
|
7
|
+
|
|
8
|
+
Use Tapp as the app's hands and eyes. Work on the real UI surface and show evidence; do not claim a
|
|
9
|
+
screen or journey works from source inspection alone.
|
|
10
|
+
|
|
11
|
+
## Choose the smallest operation
|
|
12
|
+
|
|
13
|
+
| Intent | Operation |
|
|
14
|
+
|---|---|
|
|
15
|
+
| See or screenshot one screen | `open` / `tapp_open_app` |
|
|
16
|
+
| Inspect controls on the current screen | `tree` / `tapp_ui_tree` |
|
|
17
|
+
| Drive a specific journey | MCP session start → act → end |
|
|
18
|
+
| Find bugs autonomously | `explore` / `tapp_explore` |
|
|
19
|
+
| Preserve a journey | record and save a Flow; replay it deterministically |
|
|
20
|
+
| Decide whether a merge passes policy | `ci`; exploration never decides this |
|
|
21
|
+
|
|
22
|
+
Prefer connected `tapp_*` MCP tools when available: they keep interactive sessions alive and return
|
|
23
|
+
screenshots inline. Otherwise run `npx -y @aarwitz/tapp@latest` from the app repository. Do not require MCP,
|
|
24
|
+
an account, an API key, or a global install for the core workflow.
|
|
25
|
+
|
|
26
|
+
## Start source-connected
|
|
27
|
+
|
|
28
|
+
When the user asks for a general first test of a repository:
|
|
29
|
+
|
|
30
|
+
1. If `.tapp/application-model.json` exists, run `npx -y @aarwitz/tapp@latest explore`.
|
|
31
|
+
2. Otherwise run `npx -y @aarwitz/tapp@latest init . --explore` or call `tapp_init` with
|
|
32
|
+
`{operation:"explore", projectDir:"."}`.
|
|
33
|
+
3. If Tapp returns `target-selection-required`, present its actual choices and ask the user to pick.
|
|
34
|
+
Never guess among multiple targets. Re-run with the selected platform/target exactly as Tapp
|
|
35
|
+
instructs.
|
|
36
|
+
4. If a prerequisite is missing, call `tapp_health` when MCP is connected or run
|
|
37
|
+
`npx -y @aarwitz/tapp@latest doctor`, apply only the stated remediation that is in scope, and
|
|
38
|
+
retry once.
|
|
39
|
+
|
|
40
|
+
For a focused request, use the requested target directly rather than forcing repository onboarding.
|
|
41
|
+
Targets may be a repository path, Xcode container, `.app`, iOS bundle id, APK plus Android app id,
|
|
42
|
+
or owned HTTP(S) URL. Never explore a third-party web property without authorization: exploration
|
|
43
|
+
clicks and types.
|
|
44
|
+
|
|
45
|
+
## Observe honestly
|
|
46
|
+
|
|
47
|
+
Exploration returns findings, coverage, evidence, and `inconclusive`; it does not return a score or
|
|
48
|
+
ship verdict. Report:
|
|
49
|
+
|
|
50
|
+
- target and platform;
|
|
51
|
+
- screens/actions and whether coverage was conclusive;
|
|
52
|
+
- deterministic versus advisory finding counts;
|
|
53
|
+
- each important finding and its evidence/report path;
|
|
54
|
+
- what Tapp explicitly did not check.
|
|
55
|
+
|
|
56
|
+
If `inconclusive: true`, explain the blocker. A login wall or missing test data is not a pass. Ask for
|
|
57
|
+
credentials or launch configuration instead of rerunning blindly. Do not infer content accuracy,
|
|
58
|
+
privacy, brand consistency, or business guarantees from a generic crawl; those require a reviewed
|
|
59
|
+
Flow, Scenario, contract, verifier, or human review.
|
|
60
|
+
|
|
61
|
+
When a screenshot path is printed, open it with the client's image-reading tool before describing
|
|
62
|
+
the screen. For web, use `--watch` when the human wants to follow Tapp's controlled browser. For iOS,
|
|
63
|
+
point the human to the report's exploration recording when available.
|
|
64
|
+
|
|
65
|
+
## Drive safely
|
|
66
|
+
|
|
67
|
+
For an interactive MCP session, read returned `elements[]` before every action, target accessibility
|
|
68
|
+
ids or visible labels, check `hittable`, tap a field before typing, and wait for navigation or async
|
|
69
|
+
content. Use coordinates only as a last resort. End the session when finished.
|
|
70
|
+
|
|
71
|
+
Do not edit the app merely because testing found a defect unless the user also asked for a fix. State
|
|
72
|
+
what the evidence proves and what remains untested.
|
|
73
|
+
|
|
74
|
+
Read [references/commands.md](references/commands.md) only when exact CLI/MCP syntax, Flow replay,
|
|
75
|
+
credentials, or platform prerequisites are needed.
|
|
@@ -0,0 +1,102 @@
|
|
|
1
|
+
# Tapp command reference
|
|
2
|
+
|
|
3
|
+
## Plain CLI
|
|
4
|
+
|
|
5
|
+
Run from the application repository. `[target]` is optional when Tapp can read the repository model
|
|
6
|
+
or detect one unambiguous target.
|
|
7
|
+
|
|
8
|
+
```bash
|
|
9
|
+
npx -y @aarwitz/tapp@latest init . --explore
|
|
10
|
+
npx -y @aarwitz/tapp@latest explore [target]
|
|
11
|
+
npx -y @aarwitz/tapp@latest open [target]
|
|
12
|
+
npx -y @aarwitz/tapp@latest tree [target] --json
|
|
13
|
+
npx -y @aarwitz/tapp@latest shot
|
|
14
|
+
npx -y @aarwitz/tapp@latest report latest
|
|
15
|
+
npx -y @aarwitz/tapp@latest doctor
|
|
16
|
+
```
|
|
17
|
+
|
|
18
|
+
Platform examples:
|
|
19
|
+
|
|
20
|
+
```bash
|
|
21
|
+
# Web: Tapp may build/start/stop a source target; --watch is human-visible.
|
|
22
|
+
npx -y @aarwitz/tapp@latest explore --platform web --target website --watch
|
|
23
|
+
npx -y @aarwitz/tapp@latest explore https://staging.example.com
|
|
24
|
+
|
|
25
|
+
# iOS: source repo, .app, or bundle id.
|
|
26
|
+
npx -y @aarwitz/tapp@latest explore MyApp.xcodeproj --platform ios
|
|
27
|
+
npx -y @aarwitz/tapp@latest open com.example.MyApp --platform ios
|
|
28
|
+
|
|
29
|
+
# Android: app id is required; APK is optional if already installed.
|
|
30
|
+
npx -y @aarwitz/tapp@latest explore app-debug.apk --platform android --app-id com.example.app
|
|
31
|
+
```
|
|
32
|
+
|
|
33
|
+
Focused web evidence can perform one semantic interaction and wait for async content:
|
|
34
|
+
|
|
35
|
+
```bash
|
|
36
|
+
npx -y @aarwitz/tapp@latest open https://example.com --tap "Not now" --wait-for "Dashboard"
|
|
37
|
+
```
|
|
38
|
+
|
|
39
|
+
## MCP mapping
|
|
40
|
+
|
|
41
|
+
- `tapp_init`: inspect or initialize a source repository; `operation:"explore"` prepares and explores.
|
|
42
|
+
- `tapp_build`: build and install an iOS app without needing its bundle id first.
|
|
43
|
+
- `tapp_open_app`: launch and return a screen summary plus inline screenshot.
|
|
44
|
+
- `tapp_ui_tree` / `tapp_screenshot`: inspect the current real surface.
|
|
45
|
+
- `tapp_session_start` → `tapp_session_act` → `tapp_session_end`: drive one persistent journey.
|
|
46
|
+
- `tapp_explore`: autonomous iOS, Android, or web exploration; observation only.
|
|
47
|
+
- `tapp_flow_save` / `tapp_flow_run`: save a driven journey and replay it deterministically.
|
|
48
|
+
- `tapp_release_contract`: validate, compile, or run a reviewed business guarantee.
|
|
49
|
+
- `tapp_ci_setup`: create a target-scoped baseline or reviewable CI installation.
|
|
50
|
+
|
|
51
|
+
An iOS no-bundle-id MCP path is `tapp_build {projectDir:"."}` followed by `tapp_explore` with the
|
|
52
|
+
returned `bundleId`. Android uses `androidAppId` and optional `apkPath`; web uses `url` or
|
|
53
|
+
source-connected `tapp_init`.
|
|
54
|
+
|
|
55
|
+
## Interactive session loop
|
|
56
|
+
|
|
57
|
+
```text
|
|
58
|
+
tapp_session_start {appBundleId:"com.example.app"}
|
|
59
|
+
tapp_session_act {action:"tap", id:"Email"}
|
|
60
|
+
tapp_session_act {action:"type", text:"qa@example.com"}
|
|
61
|
+
tapp_session_act {action:"tap", id:"Password"}
|
|
62
|
+
tapp_session_act {action:"type", text:"..."}
|
|
63
|
+
tapp_session_act {action:"tap", id:"Sign In"}
|
|
64
|
+
tapp_session_act {action:"wait", text:"Home", timeoutMs:10000}
|
|
65
|
+
tapp_screenshot
|
|
66
|
+
tapp_session_end
|
|
67
|
+
```
|
|
68
|
+
|
|
69
|
+
Android sessions use `androidAppId`, optional `apkPath`, and the same action loop.
|
|
70
|
+
|
|
71
|
+
## Credentials and test configuration
|
|
72
|
+
|
|
73
|
+
For autonomous exploration, pass test-only values when authorized:
|
|
74
|
+
|
|
75
|
+
- CLI: `--email`, `--password`, repeated `--launch-arg`, and JSON `--launch-env`.
|
|
76
|
+
- MCP: `testEmail`, `testPassword`, `inputOverrides`, `appLaunchArgs`, `appLaunchEnv`, or explicit
|
|
77
|
+
`loginSteps`.
|
|
78
|
+
|
|
79
|
+
Do not persist secrets in `.tapp/`. If the result reports input fields and no values were supplied,
|
|
80
|
+
ask the user rather than pretending the explored surface was complete.
|
|
81
|
+
|
|
82
|
+
## Replay and gating
|
|
83
|
+
|
|
84
|
+
Flow YAML belongs under `.tapp/flows/` and can replay without a model or API key:
|
|
85
|
+
|
|
86
|
+
```bash
|
|
87
|
+
npx -y @aarwitz/tapp@latest flow run .tapp/flows/smoke.yml
|
|
88
|
+
npx -y @aarwitz/tapp@latest ci
|
|
89
|
+
```
|
|
90
|
+
|
|
91
|
+
`tapp explore` observes. `tapp ci` applies versioned deterministic policy to evidence, selected
|
|
92
|
+
Flows/Scenarios/contracts, coverage, and any target-scoped baseline. Its outcomes are `pass`, `fail`,
|
|
93
|
+
or `inconclusive`; both `fail` and `inconclusive` block a merge.
|
|
94
|
+
|
|
95
|
+
## Platform prerequisites
|
|
96
|
+
|
|
97
|
+
- iOS: macOS, Xcode, and a booted simulator. First use builds a cached harness under `~/.tapp`.
|
|
98
|
+
- Android: `adb` and a connected authorized emulator/device.
|
|
99
|
+
- Web: Playwright and Chromium. If Tapp reports the browser missing, run
|
|
100
|
+
`npx playwright install chromium` and retry.
|
|
101
|
+
|
|
102
|
+
Captures and screenshots are stored under `~/.tapp/captures/` and `~/.tapp/shots/`.
|