code-ai-installer 1.3.0 → 1.4.0

package/locales/en/.agents/skills/qa_regression_baseline/SKILL.md

@@ -0,0 +1,217 @@
+---
+name: qa_regression_baseline
+description: >
+  Run regression tests against a saved baseline and report differences. Use this skill whenever
+  the user wants to: run regression tests, compare test results to a previous run, detect regressions
+  in code or software, update a baseline after verified changes, generate a diff report between
+  test runs, or check if recent changes broke anything. Trigger on phrases like "запусти регрессию",
+  "сравни с baseline", "проверь регрессии", "обнови baseline", "run regression", "compare with baseline",
+  "check for regressions", "regression report", "что сломалось", or any similar intent to test code
+  stability against a known-good reference point.
+---
+
+# QA Regression Baseline Skill
+
+This skill helps run regression tests, save baseline results, and detect regressions by comparing the current run with a saved standard.
+
+---
+
+## Concept
+
+```
+Baseline (standard)      Current run
+      │                       │
+      └──────── diff ──────────┘
+                  │
+            Regressions / improvements / no changes
+```
+
+- **Baseline** — saved results of a previous "good" run
+- **Regression** — a test that previously passed, but now fails (or got worse)
+- **Improvement** — a test that previously failed, but now passes
+
+---
+
+## Workflow
+
+### Step 1. Understand the context
+
+First, find out from the user (or from the conversation context):
+
+1. **What are we testing?** — project path, language/framework, command to run tests
+2. **Is there a baseline already?** — if not, suggest creating one
+3. **Results format** — JSON, JUnit XML, text logs, snapshots
+4. **What to do with discrepancies?** — report only / block / update baseline
+
+If the user didn't specify the format — determine it yourself by file extensions or log structure.
+
+---
+
+### Step 2. Determine baseline format
+
+Supported formats (read `references/formats.md` if necessary):
+
+| Format | Typical files | Frameworks |
+|--------|---------------|------------|
+| JSON | `baseline.json`, `results.json` | Universal |
+| JUnit XML | `*.xml` in `test-results/` | pytest, Jest, Maven |
+| Snapshot | `__snapshots__/*.snap` | Jest, Vitest |
+| Plain text | `*.log`, `*.txt` | Any |
+| CSV | `results.csv` | Custom scripts |
+
+---
+
+### Step 3. Run tests
+
+If the user asks to run tests — use bash:
+
+```bash
+# Python / pytest
+pytest --tb=short -q 2>&1 | tee current_run.log
+pytest --junit-xml=current_results.xml
+
+# Node.js / Jest
+npx jest --json --outputFile=current_results.json 2>&1
+
+# Universal — run whatever the user specified
+<command from config or user message>
+```
+
+Save the output to a file `current_run_<timestamp>.log` or `current_results.json`.
+
+---
+
+### Step 4. Compare with baseline
+
+Use the `scripts/compare.py` script for comparison. If the baseline does not exist — suggest creating it from the current run.
+
+```bash
+python scripts/compare.py \
+  --baseline baseline/results.json \
+  --current current_results.json \
+  --format json \
+  --output report.md
+```
+
+The script can work with JSON, JUnit XML, and plain text logs.
+
+**If baseline does not exist:**
+```bash
+python scripts/compare.py --create-baseline \
+  --source current_results.json \
+  --output baseline/results.json
+```
+
+---
+
+### Step 5. Generate report
+
+The report should contain:
+
+```
+## Regression Report — <date/time>
+
+### Summary
+- ✅ Passed: N
+- ❌ Failed: N
+- 🔴 Regressions (was OK → now FAIL): N
+- 🟢 Fixed (was FAIL → now OK): N
+- ➖ No changes: N
+
+### 🔴 Regressions (critical!)
+| Test | Status before | Status now | Details |
+|------|--------------|--------------|--------|
+| test_login | PASS | FAIL | AssertionError: ... |
+
+### 🟢 Fixed tests
+| Test | Before | Now |
+|------|--------|--------|
+
+### ℹ️ New tests (not in baseline)
+| Test | Status |
+|------|--------|
+```
+
+---
+
+### Step 6. Decide what to do with the baseline
+
+After analysis, suggest to the user:
+
+- **If no regressions** → suggest updating baseline (`--update-baseline`)
+- **If there are regressions** → show detailed diff, do not update automatically
+- **If user confirmed changes** → update baseline with a comment
+
+```bash
+# Update baseline (only after explicit user confirmation)
+cp current_results.json baseline/results.json
+echo "$(date): Updated baseline — <reason>" >> baseline/CHANGELOG.md
+```
+
+---
+
+## Project File Structure
+
+Recommended structure for storing baselines in a project:
+
+```
+project/
+├── baseline/
+│   ├── results.json        # Reference results
+│   ├── CHANGELOG.md        # Baseline update history
+│   └── config.json         # Configuration (what to compare, tolerances)
+├── test-results/
+│   └── current_run.json    # Results of the latest run
+└── qa_config.yaml          # Skill config (optional)
+```
+
+---
+
+## Configuration File (qa_config.yaml)
+
+If there's a `qa_config.yaml` in the project — read it first:
+
+```yaml
+test_command: "pytest --junit-xml=test-results/current.xml"
+baseline_path: "baseline/results.json"
+result_format: "junit_xml"   # json | junit_xml | snapshot | text
+on_regression: "report"      # report | fail | notify
+ignore_tests:                # tests that are always flaky
+  - "test_flaky_network"
+tolerance:                   # allowed deviations (for performance tests)
+  duration_ms: 200
+```
+
+---
+
+## Special Cases
+
+### Flaky tests
+If a test fails irregularly — add it to `ignore_tests` in the config or mark it in the report as "flaky, non-blocking".
+
+### Performance regressions
+If tests include time measurements — compare with a tolerance. Read `references/performance.md` for details.
+
+### Snapshot tests (Jest/Vitest)
+When updating snapshots, use `jest --updateSnapshot` only after reviewing the diff. Never update automatically without confirmation.
+
+---
+
+## Common user commands and what to do
+
+| Phrase | Action |
+|-------|---------|
+| "запусти регрессию" (run regression) | Steps 3 → 4 → 5 |
+| "создай baseline" (create baseline) | Run tests, save as baseline |
+| "обнови baseline" (update baseline) | Only after explicit confirmation, Step 6 |
+| "покажи что сломалось" (show what broke) | Step 4 → 5, focus on regressions |
+| "сравни прогоны" (compare runs) | Step 4 with two passed files |
+| "есть ли регрессии?" (any regressions?) | Steps 3 → 4 → 5 |
+
+---
+
+## Links to additional materials
+
+- `references/formats.md` — detailed format schemas (JSON, JUnit XML, snapshot)
+- `references/performance.md` — comparison of performance metrics with tolerances
+- `scripts/compare.py` — main comparison script (read before running if customization is needed)

package/locales/en/.agents/skills/qa_security_smoke_tests/SKILL.md

@@ -0,0 +1,188 @@
+---
+name: qa_security_smoke_tests
+description: Basic security smoke tests — auth/authz, validation, leaks in errors/logs, rate limiting (if enabled), SSRF/XSS where relevant.
+---
+
+# Skill: QA Security Smoke Tests
+
+Run security smoke: auth, validation, error exposure, rate limiting.
+
+**Sections:**
+1. [Workflow](#1-workflow)
+2. [Test Cases](#2-test-cases)
+3. [Curl Commands](#3-curl)
+4. [Log Inspection](#4-logs)
+5. [Output Template](#5-output)
+
+---
+
+## 1. Workflow
+
+```
+1. Identify protected endpoints (from $api_contracts)
+2. For each protected endpoint:
+   a. Request without auth → expect 401
+   b. Request with someone else's auth → expect 403/404
+   c. Invalid input → expect 400 (not 500!)
+   d. Check error response: no stack/DB/PII
+3. For public endpoints:
+   a. Check rate limiting (if enabled)
+4. Check logs: no PII/secrets
+5. Fill results table
+```
+
+---
+
+## 2. Test Cases
+
+| TC-ID | Category | Test | Request | Expected | Priority |
+|-------|----------|------|---------|----------|:--------:|
+| SEC-TC-01 | **Auth** | No token on protected endpoint | POST /coupons (no Authorization) | 401 | P0 |
+| SEC-TC-02 | **Auth** | Invalid token | POST /coupons (garbage token) | 401 | P0 |
+| SEC-TC-03 | **Auth** | Expired token | POST /coupons (expired JWT) | 401 | P1 |
+| SEC-TC-04 | **IDOR** | Access other user's data | GET /settings/other-app-id | 403 or 404 | P0 |
+| SEC-TC-05 | **IDOR** | Modify other user's data | PUT /settings/other-app-id | 403 or 404 | P0 |
+| SEC-TC-06 | **IDOR** | Delete other user's data | DELETE /coupons/other-user-coupon | 403 or 404 | P0 |
+| SEC-TC-07 | **Validation** | Empty body | POST /coupons `{}` | 400 (not 500!) | P0 |
+| SEC-TC-08 | **Validation** | Malicious field types | POST /coupons `{"code": 12345}` | 400 | P0 |
+| SEC-TC-09 | **Validation** | Oversized input | POST /coupons `{"code": "A".repeat(10000)}` | 400 | P1 |
+| SEC-TC-10 | **NoSQL** | Operator injection | GET /coupons?code[$gt]= | 400 (not data leak!) | P0 |
+| SEC-TC-11 | **Error leak** | Trigger 500 | Cause server error | No stack/DB in response | P0 |
+| SEC-TC-12 | **Error leak** | Check error format | Any 4xx/5xx | `{ error: { code, message } }` only | P0 |
+| SEC-TC-13 | **Rate limit** | Flood endpoint | 100+ requests/min | 429 (if enabled) | P1 |
+| SEC-TC-14 | **Headers** | Check security headers | Any response | X-Content-Type-Options, etc. | P1 |
+| SEC-TC-15 | **XSS** | Script in input | POST /settings `{"title": "<script>alert(1)</script>"}` | Stored sanitized or rejected | P0 |
+
+---
+
+## 3. Curl Commands
+
+### No auth (SEC-TC-01)
+
+```bash
+curl -s -o /dev/null -w "%{http_code}" \
+  -X POST https://localhost:5173/api/v1/coupons \
+  -H "Content-Type: application/json" \
+  -d '{"code":"TEST","discount":10}'
+# Expected: 401
+```
+
+### IDOR check (SEC-TC-04)
+
+```bash
+curl -s -w "\n%{http_code}" \
+  -X GET https://localhost:5173/api/v1/settings/OTHER_APP_INSTANCE_ID \
+  -H "Authorization: Bearer <my_token>"
+# Expected: 403 or 404 (NOT 200 with other user's data!)
+```
+
+### NoSQL injection (SEC-TC-10)
+
+```bash
+curl -s -w "\n%{http_code}" \
+  "https://localhost:5173/api/v1/coupons?code[\$gt]="
+# Expected: 400 (NOT 200 with all coupons!)
+```
+
+### Error exposure check (SEC-TC-11)
+
+```bash
+curl -s -X POST https://localhost:5173/api/v1/coupons \
+  -H "Content-Type: application/json" \
+  -H "Authorization: Bearer <token>" \
+  -d '{"code": null}' | python3 -m json.tool
+# Check: no "stack", no "MongoError", no file paths in response
+```
+
+### XSS in input (SEC-TC-15)
+
+```bash
+curl -s -X PUT https://localhost:5173/api/v1/settings/MY_ID \
+  -H "Content-Type: application/json" \
+  -H "Authorization: Bearer <token>" \
+  -d '{"title": "<script>alert(1)</script>"}'
+# Check: title should be sanitized/escaped, NOT stored as raw HTML
+```
+
+### Security headers (SEC-TC-14)
+
+```bash
+curl -sI https://localhost:5173/api/v1/health/live | grep -iE "x-content-type|x-frame|strict-transport|referrer-policy"
+# Expected: security headers present
+```
+
+---
+
+## 4. Log Inspection
+
+### After running tests, check logs for PII/secrets
+
+```bash
+# Check application logs (Docker)
+docker compose logs api --tail 100 | grep -iE "password|token|secret|apiKey|authorization"
+# Expected: ZERO matches (all should be [REDACTED])
+
+# Check for stack traces in response bodies
+docker compose logs api --tail 100 | grep -i "stack"
+# Expected: internal logs only, never in HTTP responses
+
+# Check for PII
+docker compose logs api --tail 100 | grep -iE "email.*@|phone.*\d{7}"
+# Expected: masked or absent
+```
+
+---
+
+## 5. Output Template
+
+```markdown
+# Security Smoke Test Report
+
+**Date:** YYYY-MM-DD
+**Tester:** Tester Agent
+**Environment:** <URL>
+
+## Results
+
+| TC-ID | Category | Test | Expected | Actual | Status |
+|-------|----------|------|:--------:|:------:|:------:|
+| SEC-TC-01 | Auth | No token | 401 | 401 | ✅ |
+| SEC-TC-04 | IDOR | Other user's data | 403/404 | 200 😱 | ❌ |
+| SEC-TC-07 | Validation | Empty body | 400 | 500 | ❌ |
+| SEC-TC-10 | NoSQL | Operator injection | 400 | 200 | ❌ |
+| SEC-TC-11 | Error leak | Trigger 500 | No stack | Has stack | ❌ |
+| SEC-TC-14 | Headers | Security headers | Present | Missing | ⚠️ |
+
+## Log Inspection
+| Check | Result |
+|-------|--------|
+| Secrets in logs | ✅ Clean |
+| Stack traces in logs | ✅ Internal only |
+| PII in logs | ⚠️ Email unmasked |
+
+## Summary
+| Status | Count |
+|--------|:-----:|
+| ✅ PASS | X |
+| ❌ FAIL | Y |
+| ⚠️ WARN | Z |
+
+## Critical Findings
+| # | Severity | Finding | Impact | Fix |
+|---|----------|---------|--------|-----|
+| 1 | 🔴 P0 | IDOR: other user's settings accessible | Data breach | Add ownership check |
+| 2 | 🔴 P0 | NoSQL injection: $gt operator not filtered | Data leak | Add input sanitization |
+
+## Verdict
+- ✅ SECURE — all security smoke tests pass
+- ⚠️ ISSUES — P1 findings to address
+- ❌ VULNERABLE — P0 security findings, block release
+```
+
+---
+
+## See also
+- `$security_review` — deep AppSec review (code level)
+- `$security_review_baseline` — quick security checklist
+- `$qa_api_contract_tests` — functional API tests
+- `$qa_test_plan` — test plan (references this skill)

package/locales/en/.agents/skills/qa_test_plan/SKILL.md

@@ -0,0 +1,203 @@
+---
+name: qa_test_plan
+description: Create a test plan and coverage matrix based on PRD/UX — roles, critical user flows, UI states, negative cases, acceptance criteria.
+---
+
+# Skill: QA Test Plan
+
+Test plan with coverage matrix based on PRD/UX flows.
+
+**Sections:**
+1. [Workflow](#1-workflow)
+2. [Test Plan Structure](#2-structure)
+3. [Coverage Matrix](#3-matrix)
+4. [Risk-Based Prioritization](#4-priority)
+5. [Entry/Exit Criteria](#5-criteria)
+6. [Example](#6-example)
+7. [Output Template](#7-output)
+
+---
+
+## 1. Workflow
+
+```
+1. Read PRD / UX Spec / Architecture Doc
+   └── Extract user flows and acceptance criteria
+
+2. Determine scope:
+   ├── What we test (in scope)
+   └── What we DO NOT test (out of scope + why)
+
+3. Break down into test suites by categories:
+   ├── Functional (happy + edge + error paths)
+   ├── API contracts
+   ├── Security smoke
+   ├── UI/a11y smoke
+   └── Performance baseline
+
+4. For each flow → test cases (section 3)
+
+5. Prioritization (section 4)
+
+6. Determine entry/exit criteria (section 5)
+```
+
+---
+
+## 2. Test Plan Structure
+
+| Section | Content |
+|---------|---------|
+| **Scope** | What's included, what's excluded (with rationale) |
+| **Assumptions** | Prerequisites, test environment conditions |
+| **Test suites** | Groups of test cases by feature/area |
+| **Coverage matrix** | Flow → test cases mapping |
+| **Test data** | Required data setup / seed / fixtures |
+| **Environment** | URLs, credentials, infra requirements |
+| **Blockers/Risks** | What could prevent testing |
+| **Schedule** | When each suite runs (CI / manual / release) |
+
+---
+
+## 3. Coverage Matrix
+
+### Test case format
+
+| Field | Description | Example |
+|-------|-------------|---------|
+| **TC-ID** | Unique ID | TC-COUPON-01 |
+| **Flow** | User flow being tested | Create coupon |
+| **Type** | happy / edge / error / security | happy |
+| **Preconditions** | Required state | User logged in, app installed |
+| **Steps** | Numbered actions | 1. Open settings 2. Fill form 3. Click Save |
+| **Expected Result** | What should happen | Coupon saved, toast shown |
+| **Priority** | P0 / P1 / P2 | P0 |
+| **Automation** | Unit / Integration / E2E / Manual | Integration |
+
+### Coverage matrix example
+
+| Flow | Happy | Edge | Error | Security | Total |
+|------|:-----:|:----:|:-----:|:--------:|:-----:|
+| Install webhook | TC-01 | TC-02 | TC-03, TC-04 | TC-05 | 5 |
+| Settings CRUD | TC-06 | TC-07, TC-08 | TC-09, TC-10 | TC-11 | 6 |
+| Coupon create | TC-12 | TC-13, TC-14 | TC-15, TC-16, TC-17 | TC-18 | 7 |
+| Widget render | TC-19 | TC-20 | TC-21 | TC-22 | 4 |
+| **Total** | 4 | 6 | 7 | 4 | **22** |
+
+---
+
+## 4. Risk-Based Prioritization
+
+| Priority | Criteria | Example |
+|----------|---------|---------|
+| 🔴 **P0** | Core functionality, data loss, security | Auth flow, payment, CRUD on main entities |
+| 🟠 **P1** | Important but recoverable, edge cases | Duplicate handling, pagination, error states |
+| 🟡 **P2** | Nice-to-have, cosmetic, rare paths | Tooltip text, animation timing, unusual screen sizes |
+
+### Priority distribution guideline
+
+| Priority | % of total test cases |
+|----------|:-----:|
+| P0 | 40-50% |
+| P1 | 30-40% |
+| P2 | 10-20% |
+
+---
+
+## 5. Entry/Exit Criteria
+
+### Entry criteria (start testing when)
+
+| # | Criterion | Status |
+|---|----------|--------|
+| 1 | Build successful | ☐ |
+| 2 | CI lint + unit tests pass | ☐ |
+| 3 | Environment accessible (dev/staging) | ☐ |
+| 4 | Test data seeded | ☐ |
+| 5 | All P0 dev tasks completed | ☐ |
+
+### Exit criteria (release when)
+
+| # | Criterion | Status |
+|---|----------|--------|
+| 1 | All P0 test cases PASS | ☐ |
+| 2 | All P1 test cases PASS (or waiver) | ☐ |
+| 3 | No open P0 bugs | ☐ |
+| 4 | Security smoke PASS | ☐ |
+| 5 | API contract compliance PASS | ☐ |
+
+---
+
+## 6. Example
+
+```markdown
+# Test Plan: SaaS Admin Panel — v1
+
+## Scope
+- IN: Settings CRUD, Product CRUD, Public content rendering
+- OUT: Analytics, multi-language (not in MVP)
+
+## Test Suites
+1. API Contract Tests (8 cases) → $qa_api_contract_tests
+2. Manual Functional (12 cases) → $qa_manual_run
+3. Security Smoke (6 cases) → $qa_security_smoke_tests
+4. UI/a11y Smoke (4 cases) → $qa_ui_a11y_smoke
+
+## Coverage: 30 test cases (P0: 12, P1: 13, P2: 5)
+## Entry: CI green, staging up, test data seeded
+## Exit: All P0 pass, no open P0 bugs
+```
+
+---
+
+## 7. Output Template
+
+```markdown
+# QA Test Plan: <feature/sprint>
+
+**Date:** YYYY-MM-DD
+**PRD:** <reference>
+**UX Spec:** <reference>
+
+## Scope
+- **In scope:** <list>
+- **Out of scope:** <list + rationale>
+
+## Assumptions
+- <list>
+
+## Test Suites
+| Suite | Cases | Priority mix | Method |
+|-------|:-----:|-------------|--------|
+| Functional | 15 | 6×P0, 6×P1, 3×P2 | Manual + Unit |
+| API Contract | 8 | 4×P0, 4×P1 | Integration |
+| Security Smoke | 6 | 4×P0, 2×P1 | Manual + Grep |
+| UI/a11y | 4 | 1×P0, 2×P1, 1×P2 | Manual |
+
+## Coverage Matrix
+<section 3 table>
+
+## Test Data
+- <required fixtures/seed data>
+
+## Environment
+- URL: <staging URL>
+- Credentials: <how to get>
+
+## Entry/Exit Criteria
+<section 5 tables>
+
+## Risks & Blockers
+| Risk | Impact | Mitigation |
+|------|--------|-----------|
+| Wix API down | P0: Install flow blocked | Skip, retest later |
+```
+
+---
+
+## See also
+- `$qa_manual_run` — execute manual test cases
+- `$qa_api_contract_tests` — API contract test execution
+- `$qa_security_smoke_tests` — security smoke execution
+- `$qa_ui_a11y_smoke` — UI/a11y smoke execution
+- `$qa_browser_testing` — browser-based visual testing

package/locales/en/.agents/skills/qa_test_plan/agents/claude.json

@@ -0,0 +1,18 @@
+{
+  "name": "qa_test_plan",
+  "display_name": "QA Test Plan",
+  "description": "Create a QA test plan and coverage matrix from PRD and UX - roles, critical flows, UI states, negative cases, and acceptance criteria.",
+  "default_prompt": "Use $qa_test_plan when the task matches the \"QA Test Plan\" skill.",
+  "triggers": [
+    "qa_test_plan",
+    "qa test plan",
+    "Create a QA test plan and coverage matrix from PRD and UX - roles, critical flows, UI states, negative cases, and acceptance criteria"
+  ],
+  "capabilities": [
+    "qa",
+    "test",
+    "plan"
+  ],
+  "tools": [],
+  "implicit_invocation": true
+}

package/locales/en/.agents/skills/qa_test_plan/agents/copilot.json

@@ -0,0 +1,18 @@
+{
+  "name": "qa_test_plan",
+  "display_name": "QA Test Plan",
+  "description": "Create a QA test plan and coverage matrix from PRD and UX - roles, critical flows, UI states, negative cases, and acceptance criteria.",
+  "default_prompt": "Use $qa_test_plan when the task matches the \"QA Test Plan\" skill.",
+  "triggers": [
+    "qa_test_plan",
+    "qa test plan",
+    "Create a QA test plan and coverage matrix from PRD and UX - roles, critical flows, UI states, negative cases, and acceptance criteria"
+  ],
+  "capabilities": [
+    "qa",
+    "test",
+    "plan"
+  ],
+  "tools": [],
+  "implicit_invocation": true
+}

package/locales/en/.agents/skills/qa_test_plan/agents/gemini.json

@@ -0,0 +1,18 @@
+{
+  "name": "qa_test_plan",
+  "display_name": "QA Test Plan",
+  "description": "Create a QA test plan and coverage matrix from PRD and UX - roles, critical flows, UI states, negative cases, and acceptance criteria.",
+  "default_prompt": "Use $qa_test_plan when the task matches the \"QA Test Plan\" skill.",
+  "triggers": [
+    "qa_test_plan",
+    "qa test plan",
+    "Create a QA test plan and coverage matrix from PRD and UX - roles, critical flows, UI states, negative cases, and acceptance criteria"
+  ],
+  "capabilities": [
+    "qa",
+    "test",
+    "plan"
+  ],
+  "tools": [],
+  "implicit_invocation": true
+}

package/locales/en/.agents/skills/qa_test_plan/agents/openai.yaml

@@ -0,0 +1,8 @@
+interface:
+  display_name: "QA Test Plan"
+  short_description: "Create a QA test plan and coverage matrix from PRD and UX - roles, critical flows, UI states, n…"
+  default_prompt: "Use $qa_test_plan when the task matches the \"QA Test Plan\" skill."
+dependencies:
+  tools: []
+policy:
+  allow_implicit_invocation: true

package/locales/en/.agents/skills/qa_test_plan/agents/qwen.json

@@ -0,0 +1,18 @@
+{
+  "name": "qa_test_plan",
+  "display_name": "QA Test Plan",
+  "description": "Create a QA test plan and coverage matrix from PRD and UX - roles, critical flows, UI states, negative cases, and acceptance criteria.",
+  "default_prompt": "Use $qa_test_plan when the task matches the \"QA Test Plan\" skill.",
+  "triggers": [
+    "qa_test_plan",
+    "qa test plan",
+    "Create a QA test plan and coverage matrix from PRD and UX - roles, critical flows, UI states, negative cases, and acceptance criteria"
+  ],
+  "capabilities": [
+    "qa",
+    "test",
+    "plan"
+  ],
+  "tools": [],
+  "implicit_invocation": true
+}