qualiow-exploratory-testing 1.1.0 → 1.3.0

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.
Files changed (79) hide show
  1. package/README.md +177 -0
  2. package/bin/mcli +16 -0
  3. package/bin/mobile-cli.mjs +573 -0
  4. package/bin/wadb +6 -0
  5. package/bin/wk-ios +58 -0
  6. package/bin/wkeval.mjs +44 -0
  7. package/data/domains/_default.md +181 -0
  8. package/data/domains/ecommerce.md +192 -0
  9. package/data/domains/fintech.md +200 -0
  10. package/data/domains/identity.md +88 -0
  11. package/data/domains/identity.yml +138 -0
  12. package/data/domains/marketing.md +194 -0
  13. package/data/domains/saas.md +197 -0
  14. package/data/knowledge/changelog.yml +86 -0
  15. package/data/knowledge/learned-patterns.md +172 -1
  16. package/data/knowledge/manifest.yml +210 -4
  17. package/data/knowledge/releases/v0.2.0/entries/technique-contract-narrowing.yml +150 -0
  18. package/data/knowledge/releases/v0.2.0/entries/technique-test-suite-audit.yml +116 -0
  19. package/data/knowledge/releases/v0.2.0/release.yml +20 -0
  20. package/data/knowledge/releases/v0.3.0/entries/technique-functional-diff-analysis.yml +119 -0
  21. package/data/knowledge/releases/v0.3.0/entries/technique-llm-output-verification.yml +102 -0
  22. package/data/knowledge/releases/v0.3.0/entries/technique-verification-mode-selection.yml +153 -0
  23. package/data/knowledge/releases/v0.3.0/release.yml +25 -0
  24. package/data/knowledge/releases/v0.4.0/entries/technique-authenticated-api-probing.yml +137 -0
  25. package/data/knowledge/releases/v0.4.0/entries/technique-environment-fingerprinting.yml +113 -0
  26. package/data/knowledge/releases/v0.4.0/entries/technique-silent-failure-audit.yml +86 -0
  27. package/data/knowledge/releases/v0.4.0/entries/technique-ui-api-differential.yml +97 -0
  28. package/data/knowledge/releases/v0.4.0/release.yml +28 -0
  29. package/data/knowledge/releases/v0.5.0/entries/technique-config-surface-verification.yml +77 -0
  30. package/data/knowledge/releases/v0.5.0/entries/technique-derived-value-verification.yml +81 -0
  31. package/data/knowledge/releases/v0.5.0/entries/technique-expected-behaviour-specification.yml +97 -0
  32. package/data/knowledge/releases/v0.5.0/entries/technique-presentation-integrity.yml +78 -0
  33. package/data/knowledge/releases/v0.5.0/entries/technique-release-readiness-verification.yml +110 -0
  34. package/data/knowledge/releases/v0.5.0/release.yml +33 -0
  35. package/data/targets/_example-backend.yml +123 -0
  36. package/data/targets/_example-mobile-emulation.yml +52 -0
  37. package/data/targets/_example-native-mobile.yml +61 -0
  38. package/data/targets/_example-sim-android-chrome.yml +81 -0
  39. package/data/targets/_example-sim-ios-safari.yml +80 -0
  40. package/data/templates/api-probe-matrix.md +99 -0
  41. package/data/templates/coverage-map.md +14 -0
  42. package/data/templates/expected-behaviour.md +93 -0
  43. package/dist/cli/index.js +137 -5
  44. package/dist/cli/index.js.map +1 -1
  45. package/dist/index.d.ts +350 -527
  46. package/dist/index.js +144 -5
  47. package/dist/index.js.map +1 -1
  48. package/docs/BACKEND-VERIFICATION.md +277 -0
  49. package/docs/MOBILE-SETUP.md +92 -0
  50. package/package.json +26 -12
  51. package/scripts/doctor-mobile.sh +143 -0
  52. package/scripts/setup-mobile.sh +213 -0
  53. package/skills/qa-explore/phases/00-setup.md +8 -1
  54. package/skills/qa-explore/phases/01-auth.md +4 -0
  55. package/skills/qa-explore-mobile/SKILL.md +146 -0
  56. package/skills/qa-explore-mobile/phases/00-setup.md +203 -0
  57. package/skills/qa-explore-mobile/phases/01-auth.md +127 -0
  58. package/skills/qa-explore-mobile/phases/02-charter.md +70 -0
  59. package/skills/qa-explore-mobile/phases/03-discovery.md +79 -0
  60. package/skills/qa-explore-mobile/phases/04-journeys.md +49 -0
  61. package/skills/qa-explore-mobile/phases/05-features.md +72 -0
  62. package/skills/qa-explore-mobile/phases/06-edge-cases.md +157 -0
  63. package/skills/qa-explore-mobile/phases/07-reporting.md +123 -0
  64. package/skills/qa-explore-mobile/references/mobile-edges.md +117 -0
  65. package/skills/qa-explore-quick/SKILL.md +6 -0
  66. package/skills/qa-verify-backend/SKILL.md +163 -0
  67. package/skills/qa-verify-backend/phases/00-setup.md +97 -0
  68. package/skills/qa-verify-backend/phases/01-ac-decomposition.md +58 -0
  69. package/skills/qa-verify-backend/phases/02-static-review.md +71 -0
  70. package/skills/qa-verify-backend/phases/03-live-verification.md +62 -0
  71. package/skills/qa-verify-backend/phases/03b-api-verification.md +157 -0
  72. package/skills/qa-verify-backend/phases/04-e2e-trigger.md +66 -0
  73. package/skills/qa-verify-backend/phases/05-reporting.md +135 -0
  74. package/skills/qa-verify-backend/references/api-probes.md +141 -0
  75. package/skills/qa-verify-backend/references/aws-readonly-probes.md +156 -0
  76. package/skills/qa-verify-backend/references/environment-fingerprinting.md +104 -0
  77. package/skills/qa-verify-backend/references/payload-verification.md +112 -0
  78. package/skills/qa-verify-backend/references/release-readiness.md +97 -0
  79. package/skills/qa-verify-backend/references/safety-rules.md +90 -0
@@ -0,0 +1,127 @@
1
+ # Phase 1: Authentication (mode-aware)
2
+
3
+ In both modes there is **NO transferable `storage_state` you can inject** — auth happens on
4
+ the device. Pick the branch by the target's `auth.strategy`:
5
+
6
+ - `auth.strategy: in-app` → **NATIVE branch** (account creation / login inside the app).
7
+ - `auth.strategy: interactive-sso` (or any interactive web SSO) → **WEB branch** (one-time
8
+ human login the browser profile persists).
9
+ - `auth.strategy: none` → skip this phase.
10
+
11
+ Set up once per Bash block:
12
+
13
+ ```bash
14
+ export MOBILE_CLI_STATE=/tmp/<target>-state.json
15
+ MCLI="$PWD/bin/mcli" # run from the repo root
16
+ ```
17
+
18
+ ---
19
+
20
+ ## NATIVE branch — in-app login / signup
21
+
22
+ Drive the app's own auth screens via the snapshot→click→fill loop, following the steps the
23
+ target's `auth.strategy` describes. There is no shared session, so you authenticate inside
24
+ the app each run (or reuse an existing account already on the device).
25
+
26
+ ### Account creation (when the charter targets signup, or as one-time setup)
27
+
28
+ ```bash
29
+ # Optional cold start (Android): relaunch-clean wipes app data, then cold-starts.
30
+ $MCLI relaunch-clean
31
+
32
+ $MCLI snapshot # welcome / landing
33
+ $MCLI click <create-account / sign-up ref>
34
+ $MCLI snapshot # email / details screen
35
+ $MCLI fill <email input ref> "<config test_email_pattern, e.g. qa+$(date +%s)@example.com>"
36
+ $MCLI click <continue ref>
37
+ $MCLI snapshot # OTP / verification screen
38
+ ```
39
+
40
+ **OTP on staging:** if the target config records a fixed code (`auth.static_otp`), use it
41
+ instead of waiting for a real OTP. Multi-cell OTP inputs often **do not accept a single
42
+ `fill`** — tap each cell and send one digit at a time:
43
+
44
+ ```bash
45
+ $MCLI snapshot # find the per-cell OTP refs
46
+ for d in <digits of static_otp>; do
47
+ $MCLI click <next-cell-ref>
48
+ $MCLI fill <that-ref> "$d"
49
+ done
50
+ ```
51
+
52
+ ### Existing-user login (faster for post-auth flows)
53
+
54
+ ```bash
55
+ $MCLI launch # do NOT relaunch-clean — preserves the existing account
56
+ $MCLI snapshot # if on landing, tap Log In
57
+ $MCLI click <login ref>
58
+ $MCLI fill <email/username input ref> "<existing test account>"
59
+ $MCLI click <continue ref>
60
+ # OTP same as above (static_otp on staging)
61
+ ```
62
+
63
+ Generalize email/validation rules from the target config or prior findings (e.g. some backends
64
+ reject disposable-email domains, or client vs server validation diverges) — do not assume any
65
+ specific provider's quirks.
66
+
67
+ ---
68
+
69
+ ## WEB branch — one-time interactive SSO/MFA login
70
+
71
+ The browser keeps its own profile (cookies/session) on the device, persisted across `launch`
72
+ / `open-url`. Do the login **once per device per session-lifetime; a HUMAN performs the MFA
73
+ step — never script it.** This applies to any interactive IdP (Microsoft Entra, Okta,
74
+ Google Workspace, Auth0, …) — see the web target templates (`_example-sim-ios-safari.yml`,
75
+ `_example-sim-android-chrome.yml`).
76
+
77
+ ```bash
78
+ $MCLI launch # do NOT relaunch-clean — that wipes the profile
79
+ $MCLI open-url <web.start_url e.g. .../auth/login>
80
+ $MCLI snapshot # login surface / SSO button
81
+ $MCLI click <sign-in / SSO button ref>
82
+ # → hands off to the identity provider (e.g. login.microsoftonline.com, your Okta org, …).
83
+ # A HUMAN completes username + password + MFA by hand in the sim/emulator UI.
84
+ ```
85
+
86
+ Once login completes, subsequent `launch` / `open-url` / exploration reuse the session until
87
+ it expires or the profile is wiped. **Never automate the MFA challenge.** If the IdP domain
88
+ reappears mid-session, the session lapsed — hand back to a human; do not fill credentials.
89
+
90
+ ### Returning session (already logged in)
91
+
92
+ ```bash
93
+ $MCLI launch
94
+ $MCLI open-url <web.base_url>
95
+ $MCLI snapshot # expect an authenticated landing view, not login
96
+ ```
97
+
98
+ If you see the login screen or IdP domain, re-do the interactive login above.
99
+
100
+ ### Profile hygiene (WEB) — what NOT to do
101
+
102
+ - **Do not run `relaunch-clean`** on a browser target — on Android it `pm clear`s the browser and forces a fresh SSO+MFA login.
103
+ - **Do not** inject cookies / storage_state — no supported path on a real sim browser.
104
+ - Boot, don't erase, the device — erasing the sim wipes the profile.
105
+
106
+ ---
107
+
108
+ ## Verify Auth (shared)
109
+
110
+ After auth, `$MCLI snapshot` and confirm authenticated indicators are present:
111
+ - NATIVE: the app's post-login home/landing controls.
112
+ - WEB: a known post-login heading, nav, or the user's name.
113
+
114
+ If stuck on a login/IdP screen, capture the log:
115
+
116
+ ```bash
117
+ $MCLI logs --errors --since 60
118
+ ```
119
+
120
+ **Auth issues to flag as bugs (not session blockers):**
121
+ - OTP rejected even with the documented static/staging code (NATIVE).
122
+ - A valid session dropped far sooner than expected (forced re-login mid-flow).
123
+ - A broken/blank state instead of redirecting to login when the session expires.
124
+ - Login surface renders but inputs / SSO button are unreachable in the a11y tree (blocks automation AND assistive tech).
125
+ - Client-side input validation (email format, required fields) diverges from server-side behavior.
126
+
127
+ **Update progress:** mark auth phase complete with `branch: native|web`, `user: <redacted>`.
@@ -0,0 +1,70 @@
1
+ # Phase 2: Business Context & Charter (Mobile)
2
+
3
+ Follow `../../qa-explore/phases/02-charter.md` for the structure (use the app as a user,
4
+ identify journeys, risk-rank, select heuristics, write charter). The mobile additions below
5
+ apply to both modes.
6
+
7
+ ## Use the app as a real user — for ~3 min
8
+
9
+ Drive the system under test via `$MCLI snapshot` → `click` → `snapshot` loops to reach entry
10
+ points (NATIVE: navigate the app's screens; WEB: `open-url` plus in-page links). Don't test
11
+ yet — just observe what it exists to do, on a phone-sized screen, the way a real mobile user
12
+ would. Note the primary task, where money/data/trust is at stake, and what a frustrated user
13
+ on a flaky connection would feel.
14
+
15
+ Concrete examples:
16
+ - **WEB mode:** the SUT is the target's mobile web app (an SPA or responsive site used on
17
+ phones in the field) — high-value areas are typically the login/SSO flow, checkout or the
18
+ core transactional flow, settings, any onboarding wizard, and the mobile form elements.
19
+ - **NATIVE mode:** treat the installed app under test the same way — open it, walk its primary
20
+ task (onboarding, the core create/transact/save flow, settings) and note the money/data/
21
+ trust-bearing screens before risk-ranking.
22
+
23
+ ## Optional — load gathered requirements as charter input
24
+
25
+ If requirements/context were prepared (e.g. via `/qa-gather`, or the user pasted tickets /
26
+ acceptance criteria), use them as **charter inputs, not a checklist** — exploratory testing
27
+ is not test execution. Requirements tell you what was *deliberately designed*; your job is to
28
+ also find what *wasn't* (missing states, unhandled errors, mobile-only gaps). Risk-rank the
29
+ feature areas yourself based on blast radius (data loss, money, security, trust) and adjust
30
+ from any provided priority order if your observation says otherwise.
31
+
32
+ ## Heuristics — mobile-flavored picks
33
+
34
+ In addition to qa-explore's SFDIPOT picks, **always select at least 1 mobile-specific
35
+ dimension** appropriate to the feature under charter, e.g.:
36
+
37
+ - Submitting / paying / saving → **Lifecycle** (background mid-submit; airplane mode at submit — does it surface failure or hang?)
38
+ - Long forms / multi-step wizards → **Keyboard & rotation** (does the on-screen keyboard occlude the active input? does rotation reflow or break layout?)
39
+ - Auth / session-sensitive areas → **Lifecycle / storage** (background 10 min → return: still authenticated? page reloaded or restored?)
40
+ - Settings / content-heavy views → **Native a11y** (VoiceOver/TalkBack focus order, dynamic type, contrast)
41
+
42
+ ## Start recording
43
+
44
+ Replace qa-explore's `tracing-start` with a video:
45
+
46
+ ```bash
47
+ $MCLI record-start output/sessions/<session-dir>/videos/session.mp4
48
+ ```
49
+
50
+ If recording fails or is too heavy, take frequent screenshots instead. Recording is
51
+ nice-to-have, not required.
52
+
53
+ ## Charter additions (append to the qa-explore charter template)
54
+
55
+ ```markdown
56
+ ## Mobile Context
57
+ **Mode:** [native | web]
58
+ **Platform:** [android | ios]
59
+ **Device:** [device label]
60
+ **App under test:** [native: app id + version | web: browser bundle]
61
+ **Target URL:** [web mode only: base_url]
62
+ **Environment:** [preprod | staging | production]
63
+
64
+ ## Requirements Input (if provided)
65
+ **Source:** [gathered context / ticket / acceptance criteria / none — blind]
66
+ **Areas referenced:** [list]
67
+ **Areas excluded from this session (and why):** [e.g. requires hardware not on sim, biometrics, etc.]
68
+ ```
69
+
70
+ **Update progress:** charter complete + platform info recorded.
@@ -0,0 +1,79 @@
1
+ # Phase 3: Discovery & Mapping (Mobile)
2
+
3
+ Follow `../../qa-explore/phases/03-discovery.md` for goal and method (map structure,
4
+ refine risk). The substitutions below apply to both modes.
5
+
6
+ ## Map screens/views by name + nav path (not URLs)
7
+
8
+ There is no address bar to map by. Map by **screen/view name + how you reached it** (NATIVE:
9
+ the nav path through the app; WEB: the URL that `open-url`/a link loaded, recorded as a hint).
10
+ For each reachable screen/view:
11
+
12
+ ```bash
13
+ $MCLI snapshot
14
+ $MCLI screenshot output/sessions/<session-dir>/screenshots/discovery-<view>.png
15
+ $MCLI logs --errors --since 30
16
+ ```
17
+
18
+ Log each: `[DISCOVERED] <view-name> — <one-line description> — Risk: <P0/P1/P2/P3> — entry: <how-to-reach>`
19
+
20
+ ## Match content by visible text / a11y label — there is no testID catalog
21
+
22
+ `snapshot` reads the device **accessibility tree, not the DOM**, so you identify elements by
23
+ their **visible text and accessibility labels** (headings, links, buttons, inputs), not by
24
+ CSS selectors or a fixed testID map. Read the live snapshot tree on each view. When a control
25
+ has no usable accessible label, fall back to tapping its coordinates from `snapshot --full`.
26
+
27
+ ## App/browser log instead of console
28
+
29
+ Replace `playwright-cli console error` with:
30
+
31
+ ```bash
32
+ $MCLI logs --errors --since 60
33
+ ```
34
+
35
+ Caveat: this is the **process log** (logcat / `simctl spawn log show` scoped to the
36
+ app/browser process). In WEB mode it is the BROWSER's log, NOT the page's JS console, so it
37
+ catches gross failures (renderer crashes) but misses most page-level `console.error`. In
38
+ NATIVE mode it is the app's own log (crashes, native exceptions, bridge errors). Treat a clean
39
+ log as "no gross crash", not "no errors". Still watch for:
40
+
41
+ - Process crashes, native exceptions, or repeated reloads/restarts
42
+ - Any log line leaking PII or credentials (tokens, emails, secrets) — always file as a high-severity finding
43
+ - Mixed-content / TLS warnings (WEB)
44
+
45
+ ## Network — limited
46
+
47
+ `mobile-cli` does not proxy traffic in v1, and there is no network panel. If a view seems to
48
+ make a request and the result is wrong (stale data, missing item), note "needs proxy capture"
49
+ in the bug report rather than guessing. Don't invent network details.
50
+
51
+ ## What's MISSING — mobile-specific
52
+
53
+ Beyond the qa-explore "what's not here" check, ask:
54
+ - **Permission prompts (mostly NATIVE)** — does the app explain WHY before requesting camera/notifications/location/contacts? (WEB: the same for browser permission prompts.)
55
+ - **Offline state** — what happens with no network? Is there a UI for it or just a hang/blank screen?
56
+ - **Empty states** — first launch/visit, no data: are zero-item lists helpful or just blank?
57
+ - **Pull-to-refresh / reload** — present where a user would expect it?
58
+ - **Keyboard handling** — does the on-screen keyboard cover the active input? Does "Done"/return dismiss it?
59
+ - **Rotation** — does the layout reflow, or break/clip in landscape?
60
+ - **Background→foreground** — return after 10 min: still authenticated? Does state restore (WEB: page reload or restore)?
61
+ - **Touch targets** — finger-sized, or too small/desktop-sized and hard to hit?
62
+
63
+ ## iOS-specific note — a11y aggregation
64
+
65
+ On iOS, some views **aggregate child elements into one parent accessibility label** (e.g. a
66
+ form returns one big comma-joined label of several fields). When you see a single ref with a
67
+ comma-joined label of many controls, that's an iOS a11y aggregation issue — file as a Medium
68
+ bug (it blocks automation and hurts VoiceOver users) and tap by approximate coordinates from
69
+ the snapshot bounds for the rest of the session on that view.
70
+
71
+ ## After Discovery
72
+
73
+ Write findings to `phase-1-discovery.md` including:
74
+ - View map (table: view → key visible labels/controls → entry path → URL if known)
75
+ - Log error summary (count by level) with the browser-log caveat noted
76
+ - iOS a11y aggregation hits (if any)
77
+ - Updated risk ranking
78
+
79
+ **Update progress:** discovery complete with `pages_explored` (views), `app_log_errors`, `missing_checklist_items`, `bugs_so_far`.
@@ -0,0 +1,49 @@
1
+ # Phase 4: End-to-End User Journeys (Mobile)
2
+
3
+ Follow `../../qa-explore/phases/04-journeys.md` for goal and the data-integrity
4
+ discipline. Build the journey templates from the charter's top user goals. Generic
5
+ mobile journey shapes to instantiate (apply to both modes):
6
+
7
+ ## Standard mobile journeys
8
+
9
+ 1. **First-run / entry → primary task** — launch the app (NATIVE) or open the start URL (WEB), complete the main thing it exists for (submit a form, complete a wizard, make a transaction, change a setting), confirm the success state actually persists.
10
+ 2. **Create / edit / save** — create or edit a record, save it, navigate away and back, confirm the change is reflected (not just an optimistic UI that drops on reload/relaunch).
11
+ 3. **Multi-step flow / wizard** — go through a multi-step flow start to finish; then go partway, back out, and resume — does it preserve or correctly reset state?
12
+ 4. **Cross-screen consistency** — a value shown on one screen (a count, total, status, balance, name) matches everywhere else it appears.
13
+ 5. **Deep entry** — reach a deep screen directly (NATIVE: deep link via `$MCLI deep-link`; WEB: `open-url` a deep view) rather than navigating from home — does it load with auth/context intact, or break?
14
+
15
+ Concrete examples:
16
+ - **WEB:** login/SSO → landing; the core transactional journey end-to-end (e.g. search → product → basket → checkout); any onboarding wizard; editing settings and confirming persistence; exercising the mobile form elements (pickers, toggles, date/number inputs).
17
+ - **NATIVE:** onboarding/account-creation → home; the app's core create/transact/save flow with verification on a related screen; a settings change that must persist across relaunch.
18
+
19
+ ## Data integrity (MANDATORY)
20
+
21
+ After every state-changing action, verify on a related view. Generic correctness oracles:
22
+
23
+ - **After a save/edit:** the new value is shown back correctly on the detail screen AND survives a reload (NATIVE: navigate away and back, or relaunch; WEB: `open-url` the same view again).
24
+ - **After a create:** the new item appears in its list/index with the right attributes.
25
+ - **After a delete:** the item is gone from every view that listed it.
26
+ - **Cross-view:** a total/count/status equals the sum or state of its parts across views.
27
+ - **Formatting:** numbers, dates, currencies, and localized text render correctly at mobile width (no truncation, no `NaN`, no `undefined`, correct locale).
28
+
29
+ Log every check: `[DATA CHECK] <oracle> — expected: <X>, actual: <Y> — MATCH/MISMATCH`
30
+
31
+ ## Lifecycle interleaving (mobile-only addition)
32
+
33
+ For at least ONE journey, interrupt mid-flow and verify recovery:
34
+ - Background the app/browser via `$MCLI press HOME` (Android) / home gesture (iOS), wait 30s, return — does state restore, or reload-and-lose-input?
35
+ - Kill via `$MCLI stop`, relaunch via `$MCLI launch` (WEB: + `open-url`) — does in-progress work survive? Does auth survive?
36
+ - Toggle airplane mode mid-submit (`bin/wadb shell svc wifi disable; bin/wadb shell svc data disable`) — does the app surface the failure cleanly or hang/silently drop the action? Re-enable to undo.
37
+
38
+ These lifecycle results often surface the most impactful bugs (data loss, stuck states,
39
+ silent failures, duplicate submissions).
40
+
41
+ ## After Journeys
42
+
43
+ Write findings to `phase-2-user-journeys.md` with:
44
+ - Journey-by-journey result (passed / failed / partial — and the WHY)
45
+ - Data integrity table (oracle, expected, actual, pass/fail)
46
+ - Lifecycle test results
47
+ - Cross-view inconsistencies (e.g., "Step 3 shows value X, the summary shows Y")
48
+
49
+ **Update progress:** journeys complete, data integrity counts.
@@ -0,0 +1,72 @@
1
+ # Phase 5: Deep Feature Testing (Mobile)
2
+
3
+ Follow `../../qa-explore/phases/05-features.md` for the SFDIPOT structure, boundary
4
+ values, business-logic stress, and FEW HICCUPPS oracles. The mobile changes (both modes):
5
+
6
+ ## Substitutions
7
+
8
+ - **`playwright-cli eval` (parameter tampering)** — **DROPPED**. No script context in either
9
+ mode. If a test *would* benefit from tampering server-controlled values (hidden fields,
10
+ computed totals, query params), document the scenario and mark **"deferred — needs proxy
11
+ tooling (mitmproxy)"**. Do NOT invent a result. (You CAN still tamper values that travel in
12
+ a deep link / URL via `deep-link`/`open-url` — see Phase 6.)
13
+
14
+ - **`playwright-cli fill <ref> <long-string>`** — use the same patterns, but be aware the
15
+ Android `input text` command silently truncates very long strings (~5000 chars max on most
16
+ emulators), so a "field accepted N chars" result may be a driver limit, not the app's.
17
+ Note the truncation when reporting, and for true very-long tests prefer pasting via the
18
+ clipboard:
19
+ ```bash
20
+ bin/wadb shell "echo '<very long string>' | xargs -I {} am broadcast -a clipper.set -e text '{}'"
21
+ $MCLI click <input ref>
22
+ # then long-press to paste — or use $MCLI fill with the actual string and note truncation
23
+ ```
24
+
25
+ ## Generic mobile feature stress recipes
26
+
27
+ Instantiate these for the feature under test (pick the ones that fit; they apply to native
28
+ screens and web views alike — they replace the desktop-only Playwright specifics).
29
+
30
+ ### Form fields & inputs
31
+
32
+ - **Boundary values:** empty, single char, max length, max+1, leading/trailing whitespace, whitespace-only.
33
+ - **Character classes:** unicode/emoji (`naïve café 🤑 العربية`), RTL text, control chars, paste with formatting.
34
+ - **Client vs server validation:** does the field reject client-side what the server would also reject — and accept what the server actually accepts? Try values just outside the visible UI constraint.
35
+ - **Keyboard occlusion:** when the input is near the bottom, does the on-screen keyboard cover it or the submit button? Does the view scroll the input into view?
36
+ - **Keyboard type:** does a numeric/email field bring up the right mobile keyboard?
37
+
38
+ ### Pickers, toggles, steppers, date/number controls
39
+
40
+ - Min/max/step boundaries; rapid toggling; selecting then dismissing without choosing.
41
+ - Native vs custom picker behavior on a touch device — is the custom control reachable and operable by tap (and by a11y)?
42
+ - Does the chosen value persist on reload and reflect everywhere it's shown?
43
+
44
+ ### Multi-step flows / wizards
45
+
46
+ - Forward-only vs back-navigable: go back mid-wizard — is prior input preserved?
47
+ - Skip an optional step then return — consistent state?
48
+ - Submit twice / double-tap the final action — duplicate submission? Idempotency?
49
+ - Reload mid-wizard (NATIVE: background/relaunch; WEB: `open-url` the current step) — resume or lose everything?
50
+
51
+ ### Navigation & routing
52
+
53
+ - Hardware/gesture back (Android `press BACK`) vs in-screen back control — same destination?
54
+ - Deep-link straight into a mid-flow screen (NATIVE: `$MCLI deep-link <app-scheme>`; WEB: `open-url`) — does it load with context or break?
55
+ - After a successful action, does back land somewhere sane (not re-submitting)?
56
+
57
+ ### Submit / save actions
58
+
59
+ - Slow network at submit: does it show progress, prevent double-submit, and confirm success?
60
+ - Background the app/browser on a confirm/submit screen → return → submit: stale data? re-validated?
61
+
62
+ Concrete high-value targets:
63
+ - **WEB:** the **mobile form elements** (pickers, toggles, date/number inputs — do they work by touch and persist?), any **multi-step wizard/checkout** (back/resume/skip), and **settings** (do changes save and reflect app-wide?).
64
+ - **NATIVE:** the app's primary transactional/create flow under boundary + replay stress, its custom input controls, and any setting that must persist across relaunch.
65
+
66
+ ## After Feature Testing
67
+
68
+ Write `phase-3-feature-testing.md` with feature-by-feature notes. For each deferred test
69
+ (proxy-blocked, multi-device-blocked), record it explicitly in the coverage map so it shows
70
+ up as a known gap in the report.
71
+
72
+ **Update progress.**
@@ -0,0 +1,157 @@
1
+ # Phase 6: Edge Cases & Mobile-Specific (mode-aware)
2
+
3
+ Follow `../../qa-explore/phases/06-edge-cases.md` for the saboteur mindset. Edges are tagged
4
+ **[both]**, **[native]**, or **[web]** — run the ones that fit the resolved mode.
5
+
6
+ Set up once per Bash block:
7
+ ```bash
8
+ export MOBILE_CLI_STATE=/tmp/<target>-state.json
9
+ MCLI="$PWD/bin/mcli" # run from the repo root
10
+ ```
11
+
12
+ ## Dropped / limited (no DOM, no JS console)
13
+
14
+ - **XSS execution assertions via `fill` [web]** — without DOM/console you can't confirm execution. You can still submit payloads and watch for gross breakage (crash, broken render, reflected unescaped text in the a11y tree).
15
+ - **`route ... --status=500` network simulation [both]** — needs mitmproxy. **Deferred** — log scenarios that would benefit, run them once a proxy exists.
16
+ - **Multi-session race [both]** — needs two emulators. **Deferred for v1.**
17
+ - **`resize` viewport tests** — sim viewport is fixed. Use **rotation** instead (below).
18
+
19
+ ## Kept (still useful on mobile)
20
+
21
+ - **Unicode / emoji [both]** in inputs: `$MCLI fill <ref> "naïve café 🤑 العربية"`
22
+ - **Whitespace-only [both]:** `$MCLI fill <ref> " "`
23
+ - **Extremely long [both]:** see the truncation caveat in `05-features.md`
24
+ - **State manipulation via deep link / URL [both]:** NATIVE `$MCLI deep-link "<app-scheme>://<path>?<tampered>"`; WEB `$MCLI open-url "<base_url>/<deep view>?<tampered params>"`
25
+ - **Replay [both]:** double-tap submit; submit then back then re-submit
26
+ - **Brute force / rate limits [both]:** repeated failed submissions/logins — lockout / cooldown?
27
+
28
+ ## Mobile-device edges (the high-value additions)
29
+
30
+ ### Lifecycle [both]
31
+
32
+ ```bash
33
+ # Background mid-flow
34
+ $MCLI press HOME # Android
35
+ # (iOS: use the home gesture/button via the Simulator; press HOME is Android-only)
36
+ sleep 30
37
+ $MCLI launch # foreground — did state restore or reload-and-lose-state?
38
+
39
+ # Kill mid-flow
40
+ $MCLI stop
41
+ $MCLI launch
42
+ # WEB only: re-open the view you were on
43
+ # $MCLI open-url <current view>
44
+
45
+ # OS-initiated suspend (Android)
46
+ bin/wadb shell am broadcast -a android.intent.action.SCREEN_OFF
47
+ sleep 5
48
+ bin/wadb shell am broadcast -a android.intent.action.SCREEN_ON
49
+ ```
50
+
51
+ For each: did state survive? Was auth still valid? Did an in-progress action complete or
52
+ duplicate?
53
+
54
+ ### Permissions [native] (browser prompts in [web])
55
+
56
+ NATIVE — deny / revoke / re-grant runtime permissions and check graceful handling:
57
+ ```bash
58
+ bin/wadb shell pm revoke <app.package> android.permission.CAMERA # then try the QR/camera feature
59
+ bin/wadb shell pm grant <app.package> android.permission.CAMERA
60
+ # iOS sim: xcrun simctl privacy <udid> reset|grant camera <app.bundle_id>
61
+ ```
62
+ Test camera / notifications / contacts / location. In WEB mode the same applies to the
63
+ **browser** package (controls page getUserMedia/geolocation/notifications). Ask: does denial
64
+ crash or degrade gracefully? Does the app explain WHY before prompting? Can the user re-grant?
65
+
66
+ ### Biometrics [native]
67
+
68
+ If the app uses biometric auth: approve → succeeds; deny/cancel → password fallback works;
69
+ disable biometrics in OS mid-flow → graceful fallback. Android emulator can simulate a
70
+ fingerprint touch:
71
+ ```bash
72
+ bin/wadb -e emu finger touch 1
73
+ ```
74
+
75
+ ### Push notifications [native]
76
+
77
+ If the app registers for push: deny the notification permission and confirm the app still
78
+ works; trigger a notification and tap it — does it deep-link to the right screen with context?
79
+
80
+ ### Connectivity [both]
81
+
82
+ ```bash
83
+ # Android — toggle network
84
+ bin/wadb shell svc wifi disable && bin/wadb shell svc data disable
85
+ # Run a submit/save flow — does it surface "offline" cleanly or hang/blank?
86
+ bin/wadb shell svc wifi enable && bin/wadb shell svc data enable
87
+
88
+ # Slow network: use the emulator console — telnet localhost 5554 → network speed gsm / network delay umts
89
+ ```
90
+
91
+ Ask: does the app surface offline cleanly, queue or fail loudly, auto-retry, or just spin?
92
+
93
+ ### Rotation [both]
94
+
95
+ ```bash
96
+ # Android
97
+ bin/wadb shell settings put system accelerometer_rotation 0
98
+ bin/wadb shell settings put system user_rotation 1 # landscape
99
+ $MCLI snapshot # does the layout reflow? keyboard cover input?
100
+ bin/wadb shell settings put system user_rotation 0 # portrait
101
+ ```
102
+
103
+ iOS sim: Cmd+Right/Left Arrow in the Simulator to rotate. Check reflow, clipping, and that
104
+ no content becomes unreachable in landscape.
105
+
106
+ ### On-screen keyboard [both]
107
+
108
+ Focus an input near the bottom of the screen and snapshot — does the keyboard occlude the
109
+ input or the submit button? Does the view scroll it into view? Does return/"Done" dismiss it
110
+ and submit correctly?
111
+
112
+ ### Deep links / URL attack surface [both]
113
+
114
+ ```bash
115
+ # NATIVE — exercise the app's own scheme/intent:
116
+ $MCLI deep-link "<app-scheme>://<a real path>"
117
+ $MCLI deep-link "<app-scheme>://<path>?id=&n=NaN"
118
+ bin/wadb shell am start -a android.intent.action.VIEW -d "<app-scheme>://<path>?x=$(python3 -c 'print("A"*5000)')"
119
+
120
+ # WEB — exercise the target URL:
121
+ $MCLI open-url "<base_url>/<a real deep view>"
122
+ $MCLI open-url "<base_url>/<view>?id=&n=NaN"
123
+ $MCLI open-url "<base_url>/<view>?next=javascript:alert(1)"
124
+ $MCLI open-url "<base_url>/<view>?q=$(python3 -c 'print("A"*5000)')"
125
+ ```
126
+
127
+ Crash, freeze, blank screen, open-redirect, or accept-as-valid are all findings.
128
+
129
+ ### Privacy in app-switcher / Recents [both]
130
+
131
+ ```bash
132
+ # Navigate to a view with sensitive data, then open Recents and screenshot the preview
133
+ bin/wadb shell input keyevent KEYCODE_APP_SWITCH # Android Recents
134
+ $MCLI screenshot output/sessions/<session-dir>/screenshots/recents-preview.png
135
+ ```
136
+
137
+ Is sensitive data (account info, tokens-in-URL, personal data) visible in the Recents preview?
138
+ A native app can mask its preview (FLAG_SECURE); web apps generally can't — note the leak as a
139
+ finding where it matters.
140
+
141
+ ### Storage / data wipe
142
+
143
+ - **NATIVE:** force-clear app data mid-flow (Android `$MCLI relaunch-clean` = force-stop + `pm clear`) — does it land on a clean welcome state or get stuck on an intermediate screen? Also test fresh-install vs upgrade if you have both builds.
144
+ - **WEB:** **do NOT use `relaunch-clean`** — on Android it `pm clear`s the *browser* and wipes the logged-in profile, forcing a fresh SSO+MFA login. If you need a cleared-cookie test, do it knowingly and expect to re-login.
145
+ - Returning vs fresh state [both]: correct first-ever behavior (empty states, first-run interstitials) AND returning behavior.
146
+
147
+ ### Empty / boundary states [both]
148
+
149
+ For each list / feed: zero items, exactly one item, many items. Check formatters at the
150
+ extremes (very large numbers, long names, long localized strings at mobile width).
151
+
152
+ ## After Edge Cases
153
+
154
+ Write `phase-4-edge-cases.md` with one section per category. Each deferred test gets a single
155
+ line so it's visible.
156
+
157
+ **Update progress.**