clavix 4.7.0 → 4.8.0

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.
@@ -0,0 +1,292 @@
1
+ ---
2
+ name: "Clavix: Verify"
3
+ description: Verify implementation against validation checklist from deep/fast mode
4
+ ---
5
+
6
+ # Clavix: Verify Implementation
7
+
8
+ Verify that your implementation covers the validation checklist, edge cases, and risks identified by `/clavix:deep` or `/clavix:fast`.
9
+
10
+ ---
11
+
12
+ ## CLAVIX MODE: Verification
13
+
14
+ **You are in Clavix verification mode. You verify implementations against checklists.**
15
+
16
+ **YOUR ROLE:**
17
+ - ✓ Load saved prompt and extract validation checklist
18
+ - ✓ Verify each checklist item systematically
19
+ - ✓ Run automated hooks where applicable (test, build, lint)
20
+ - ✓ Report pass/fail/skip for each item with reasoning
21
+ - ✓ Generate verification report
22
+
23
+ **DO NOT IMPLEMENT. DO NOT MODIFY CODE.**
24
+ - ✗ Writing new code
25
+ - ✗ Fixing issues found during verification
26
+ - ✗ Making changes to implementation
27
+ - Only verify and report. User will fix issues and re-verify.
28
+
29
+ **MODE ENTRY VALIDATION:**
30
+ Before verifying, confirm:
31
+ 1. Prompt was executed via `/clavix:execute`
32
+ 2. Implementation is complete (or ready for verification)
33
+ 3. Output assertion: "Entering VERIFICATION mode. I will verify, not implement."
34
+
35
+ ---
36
+
37
+ ## Prerequisites
38
+
39
+ 1. Generate optimized prompt with checklist:
40
+ ```bash
41
+ /clavix:deep "your requirement"
42
+ ```
43
+
44
+ 2. Execute and implement:
45
+ ```bash
46
+ /clavix:execute --latest
47
+ # ... implement requirements ...
48
+ ```
49
+
50
+ 3. Then verify:
51
+ ```bash
52
+ /clavix:verify --latest
53
+ ```
54
+
55
+ ---
56
+
57
+ ## Usage
58
+
59
+ **Verify latest executed prompt (recommended):**
60
+ ```bash
61
+ clavix verify --latest
62
+ ```
63
+
64
+ **Verify specific prompt:**
65
+ ```bash
66
+ clavix verify --id <prompt-id>
67
+ ```
68
+
69
+ **Show verification status:**
70
+ ```bash
71
+ clavix verify --status
72
+ ```
73
+
74
+ **Re-run only failed items:**
75
+ ```bash
76
+ clavix verify --retry-failed
77
+ ```
78
+
79
+ **Export verification report:**
80
+ ```bash
81
+ clavix verify --export markdown
82
+ clavix verify --export json
83
+ ```
84
+
85
+ ---
86
+
87
+ ## Verification Workflow
88
+
89
+ ### Step 1: Load Checklist
90
+
91
+ 1. Locate saved prompt from `.clavix/outputs/prompts/deep/` or `.clavix/outputs/prompts/fast/`
92
+ 2. Extract validation checklist, edge cases, and risks sections
93
+ 3. If fast mode (no checklist): Generate basic checklist from intent
94
+
95
+ ### Step 2: Run Automated Hooks
96
+
97
+ For items that can be verified automatically:
98
+
99
+ | Hook | Verifies |
100
+ |------|----------|
101
+ | `test` | "Tests pass", "All tests", "Test coverage" |
102
+ | `build` | "Compiles", "Builds without errors" |
103
+ | `lint` | "No warnings", "Follows conventions" |
104
+ | `typecheck` | "Type errors", "TypeScript" |
105
+
106
+ Hooks are auto-detected from `package.json` scripts.
107
+
108
+ ### Step 3: Manual Verification
109
+
110
+ For each item that requires manual verification:
111
+
112
+ 1. **Display the item:**
113
+ ```
114
+ 📋 [Item description]
115
+ Category: [Functionality/Robustness/Quality/etc.]
116
+ ```
117
+
118
+ 2. **Verify against implementation:**
119
+ - Review code changes
120
+ - Test functionality
121
+ - Check edge cases
122
+
123
+ 3. **Record result:**
124
+ - ✓ **Passed**: Item is verified (with evidence)
125
+ - ✗ **Failed**: Item not covered (with reason)
126
+ - ⏭️ **Skipped**: Will verify later
127
+ - ➖ **N/A**: Does not apply to this implementation
128
+
129
+ ### Step 4: Generate Report
130
+
131
+ ```
132
+ ══════════════════════════════════════════════════════════════════════
133
+ VERIFICATION REPORT
134
+ [prompt-id]
135
+ ══════════════════════════════════════════════════════════════════════
136
+
137
+ 📋 VALIDATION CHECKLIST (X items)
138
+
139
+ ✅ [automated] Code compiles/runs without errors
140
+ Evidence: npm run build - exit code 0
141
+ Confidence: HIGH
142
+
143
+ ✅ [manual] All requirements implemented
144
+ Evidence: Login page with OAuth, callback handling
145
+ Confidence: MEDIUM
146
+
147
+ ❌ [manual] Keyboard navigation works
148
+ Status: FAILED
149
+ Reason: Tab order skips OAuth buttons
150
+ Confidence: MEDIUM
151
+
152
+ ══════════════════════════════════════════════════════════════════════
153
+ SUMMARY
154
+ ══════════════════════════════════════════════════════════════════════
155
+ Total: X items
156
+ Passed: Y (Z%)
157
+ Failed: N (requires attention)
158
+ Skipped: M
159
+
160
+ ⚠️ N item(s) require attention before marking complete
161
+ ══════════════════════════════════════════════════════════════════════
162
+ ```
163
+
164
+ ---
165
+
166
+ ## Fast Mode Handling
167
+
168
+ When verifying a fast mode prompt (no checklist):
169
+
170
+ 1. **Detect intent** from original prompt
171
+ 2. **Generate basic checklist** based on intent:
172
+ - `code-generation`: compiles, requirements met, no errors, follows conventions
173
+ - `testing`: tests pass, coverage acceptable, edge cases tested
174
+ - `debugging`: bug fixed, no regression, root cause addressed
175
+ - etc.
176
+
177
+ 3. **Display notice:**
178
+ ```
179
+ ⚠️ No checklist found (fast mode prompt)
180
+ Generating basic checklist based on intent...
181
+
182
+ 💡 For comprehensive checklists, use /clavix:deep
183
+ ```
184
+
185
+ ---
186
+
187
+ ## Verification Methods by Category
188
+
189
+ ### Functionality
190
+ - Run the implemented feature
191
+ - Check expected behavior matches requirements
192
+ - Verify all user flows complete successfully
193
+
194
+ ### Testing
195
+ - Run test suite: `npm test`
196
+ - Check coverage report
197
+ - Verify no failing tests
198
+
199
+ ### Robustness/Edge Cases
200
+ - Test with edge case inputs (empty, null, max values)
201
+ - Check error messages are user-friendly
202
+ - Verify system recovers gracefully
203
+
204
+ ### Quality
205
+ - Run linter: `npm run lint`
206
+ - Check for console errors
207
+ - Review code style
208
+
209
+ ### Security (if applicable)
210
+ - Verify authentication required where expected
211
+ - Test input sanitization
212
+ - Check sensitive data handling
213
+
214
+ ---
215
+
216
+ ## After Verification
217
+
218
+ ### If All Items Pass
219
+ ```
220
+ ✓ Verification complete!
221
+
222
+ Next steps:
223
+ /clavix:archive - Archive completed project
224
+ clavix prompts clear --executed - Cleanup prompts
225
+ ```
226
+
227
+ ### If Items Fail
228
+ ```
229
+ ⚠️ Some items require attention.
230
+
231
+ Fix issues and re-run:
232
+ clavix verify --retry-failed --id <prompt-id>
233
+ ```
234
+
235
+ ---
236
+
237
+ ## Verification Report Storage
238
+
239
+ Reports are saved alongside prompt files:
240
+ ```
241
+ .clavix/
242
+ outputs/
243
+ prompts/
244
+ deep/
245
+ deep-20250117-143022-a3f2.md # Prompt
246
+ deep-20250117-143022-a3f2.verification.json # Report
247
+ ```
248
+
249
+ ---
250
+
251
+ ## Agent Transparency (v4.8)
252
+
253
+ ### Verification Confidence Levels
254
+
255
+ | Level | Meaning | Example |
256
+ |-------|---------|---------|
257
+ | HIGH | Automated verification passed | npm test exit code 0 |
258
+ | MEDIUM | Agent verified with evidence | Code review confirmed |
259
+ | LOW | Agent verified without clear evidence | General assessment |
260
+
261
+ ### Verification Checkpoint Output
262
+
263
+ After completing verification:
264
+ ```
265
+ VERIFICATION CHECKPOINT (v4.8):
266
+ - Prompt: [id]
267
+ - Total items: [X]
268
+ - Passed: [Y] ([Z]%)
269
+ - Failed: [N]
270
+ - Status: [completed/requires-attention]
271
+ ```
272
+
273
+ ### Error Handling
274
+ {{INCLUDE:agent-protocols/error-handling.md}}
275
+
276
+ ### Decision Rules
277
+ {{INCLUDE:agent-protocols/decision-rules.md}}
278
+
279
+ ---
280
+
281
+ ## Workflow Navigation
282
+
283
+ **You are here:** Verify (Post-Implementation Verification)
284
+
285
+ **Common workflows:**
286
+ - `/clavix:execute` → **`/clavix:verify`** → Fix issues → Re-verify → `/clavix:archive`
287
+ - `/clavix:implement` → **`/clavix:verify`** → `/clavix:archive`
288
+
289
+ **Related commands:**
290
+ - `/clavix:execute` - Execute saved prompt (previous step)
291
+ - `/clavix:deep` - Comprehensive analysis with validation checklist
292
+ - `/clavix:archive` - Archive completed project (next step)
@@ -0,0 +1,184 @@
1
+ ## Verification Methods by Category
2
+
3
+ ### Functionality
4
+ **Checklist items about:** Code works, features implemented, requirements met
5
+
6
+ **Verification approach:**
7
+ 1. Run the implemented code/feature
8
+ 2. Check expected behavior matches requirements
9
+ 3. Verify all user flows complete successfully
10
+
11
+ **Commands to use:**
12
+ - Run application: `npm start`, `npm run dev`
13
+ - Execute specific feature manually
14
+ - Check output matches expected
15
+
16
+ **Evidence examples:**
17
+ - "Feature X works as specified"
18
+ - "Login flow completes successfully"
19
+ - "API returns expected response format"
20
+
21
+ ---
22
+
23
+ ### Testing
24
+ **Checklist items about:** Tests pass, coverage met, test quality
25
+
26
+ **Verification approach:**
27
+ 1. Run test suite
28
+ 2. Check coverage report
29
+ 3. Verify no failing tests
30
+
31
+ **Commands to use:**
32
+ - `npm test` or project test command
33
+ - `npm run coverage` for coverage report
34
+ - Look for test output in terminal
35
+
36
+ **Evidence examples:**
37
+ - "npm test - 47 tests passing, 0 failed"
38
+ - "Coverage: 85% (exceeds 80% threshold)"
39
+ - "All integration tests pass"
40
+
41
+ ---
42
+
43
+ ### Robustness/Edge Cases
44
+ **Checklist items about:** Error handling, edge cases, graceful degradation
45
+
46
+ **Verification approach:**
47
+ 1. Test with edge case inputs (empty, null, max values)
48
+ 2. Check error messages are user-friendly
49
+ 3. Verify system recovers gracefully
50
+
51
+ **Manual testing:**
52
+ - Input empty values
53
+ - Input invalid data types
54
+ - Test boundary conditions (min/max values)
55
+ - Test with large datasets
56
+
57
+ **Evidence examples:**
58
+ - "Empty input shows validation error"
59
+ - "Invalid email format displays helpful message"
60
+ - "System handles 10,000 records without timeout"
61
+
62
+ ---
63
+
64
+ ### Quality
65
+ **Checklist items about:** Code style, conventions, no warnings
66
+
67
+ **Verification approach:**
68
+ 1. Run linter
69
+ 2. Check for console errors
70
+ 3. Review code style
71
+
72
+ **Commands to use:**
73
+ - `npm run lint` or equivalent
74
+ - Check browser console for errors
75
+ - Review PR diff for style
76
+
77
+ **Evidence examples:**
78
+ - "npm run lint - 0 errors, 0 warnings"
79
+ - "No console errors in browser"
80
+ - "Code follows project conventions"
81
+
82
+ ---
83
+
84
+ ### Accessibility
85
+ **Checklist items about:** Keyboard navigation, screen reader, WCAG
86
+
87
+ **Verification approach:**
88
+ 1. Tab through interface
89
+ 2. Check color contrast
90
+ 3. Verify alt text on images
91
+
92
+ **Manual testing:**
93
+ - Navigate using only keyboard
94
+ - Check focus indicators visible
95
+ - Test with screen reader (optional)
96
+
97
+ **Evidence examples:**
98
+ - "All interactive elements keyboard accessible"
99
+ - "Focus order is logical"
100
+ - "Color contrast meets WCAG AA"
101
+
102
+ ---
103
+
104
+ ### Security
105
+ **Checklist items about:** Auth, input sanitization, data protection
106
+
107
+ **Verification approach:**
108
+ 1. Verify authentication required where expected
109
+ 2. Test input sanitization
110
+ 3. Check sensitive data handling
111
+
112
+ **Manual testing:**
113
+ - Try accessing protected routes without auth
114
+ - Submit potentially malicious input
115
+ - Check network tab for sensitive data exposure
116
+
117
+ **Evidence examples:**
118
+ - "Protected routes redirect to login"
119
+ - "SQL injection attempts are sanitized"
120
+ - "Passwords not logged or exposed"
121
+
122
+ ---
123
+
124
+ ### Performance
125
+ **Checklist items about:** Response times, resource usage
126
+
127
+ **Verification approach:**
128
+ 1. Check response times
129
+ 2. Monitor memory/CPU usage
130
+ 3. Test with realistic data volumes
131
+
132
+ **Tools:**
133
+ - Browser DevTools Performance tab
134
+ - Network tab for response times
135
+ - Lighthouse performance score
136
+
137
+ **Evidence examples:**
138
+ - "Page load time < 2s"
139
+ - "API response time < 200ms"
140
+ - "Memory usage stable under load"
141
+
142
+ ---
143
+
144
+ ### Documentation
145
+ **Checklist items about:** Docs updated, comments present
146
+
147
+ **Verification approach:**
148
+ 1. Check README updates
149
+ 2. Verify JSDoc/comments on complex functions
150
+ 3. Review API documentation
151
+
152
+ **Commands to use:**
153
+ - `cat README.md` - check for updates
154
+ - Review changed files for comments
155
+
156
+ **Evidence examples:**
157
+ - "README updated with new feature"
158
+ - "API endpoints documented"
159
+ - "Complex logic has explanatory comments"
160
+
161
+ ---
162
+
163
+ ## Verification Type Detection
164
+
165
+ **Automated items contain keywords:**
166
+ - compiles, builds, tests pass, lint, typecheck, no errors
167
+
168
+ **Semi-automated items contain keywords:**
169
+ - renders, displays, console errors, responsive, visual
170
+
171
+ **Manual items contain keywords:**
172
+ - requirements, edge cases, handles, correctly, properly
173
+
174
+ ---
175
+
176
+ ## Confidence Levels
177
+
178
+ | Level | When to use | Example |
179
+ |-------|-------------|---------|
180
+ | HIGH | Automated tool verification | npm test exit code 0 |
181
+ | MEDIUM | Manual verification with clear evidence | Code review shows implementation |
182
+ | LOW | General assessment without specific evidence | "Looks correct" |
183
+
184
+ Always prefer higher confidence verification when possible.
@@ -0,0 +1,204 @@
1
+ /**
2
+ * Clavix v4.8: Verification System Types
3
+ *
4
+ * Type definitions for the checklist verification system that ensures
5
+ * checklists generated by deep/fast modes are verified after implementation.
6
+ */
7
+ import { PromptIntent } from '../core/intelligence/types.js';
8
+ /**
9
+ * Category of checklist item
10
+ */
11
+ export type ChecklistCategory = 'validation' | 'edge-case' | 'risk';
12
+ /**
13
+ * How the checklist item can be verified
14
+ */
15
+ export type VerificationType = 'automated' | 'semi-automated' | 'manual';
16
+ /**
17
+ * A single item from the checklist
18
+ */
19
+ export interface ChecklistItem {
20
+ /** Unique identifier (e.g., "validation-1", "edge-case-2") */
21
+ id: string;
22
+ /** Category of the item */
23
+ category: ChecklistCategory;
24
+ /** Item description text */
25
+ content: string;
26
+ /** Optional grouping (e.g., "Functionality", "Robustness") */
27
+ group?: string;
28
+ /** How this item should be verified */
29
+ verificationType: VerificationType;
30
+ }
31
+ /**
32
+ * Parsed checklist from a prompt file
33
+ */
34
+ export interface ParsedChecklist {
35
+ /** Validation checklist items (☐ items from deep mode) */
36
+ validationItems: ChecklistItem[];
37
+ /** Edge cases to consider */
38
+ edgeCases: ChecklistItem[];
39
+ /** Risk/what could go wrong items */
40
+ risks: ChecklistItem[];
41
+ /** Whether the prompt has any checklist */
42
+ hasChecklist: boolean;
43
+ /** Total number of items across all categories */
44
+ totalItems: number;
45
+ }
46
+ /**
47
+ * Type of verification hook
48
+ */
49
+ export type HookType = 'test' | 'build' | 'lint' | 'typecheck' | 'custom';
50
+ /**
51
+ * A CLI hook for automated verification
52
+ */
53
+ export interface VerificationHook {
54
+ /** Hook identifier */
55
+ name: HookType;
56
+ /** Display name */
57
+ displayName: string;
58
+ /** Command to execute */
59
+ command: string;
60
+ /** Regex pattern to match success */
61
+ successPattern?: RegExp;
62
+ /** Regex pattern to match failure */
63
+ failurePattern?: RegExp;
64
+ /** Timeout in milliseconds */
65
+ timeout: number;
66
+ }
67
+ /**
68
+ * Result of running a verification hook
69
+ */
70
+ export interface HookResult {
71
+ /** Hook that was run */
72
+ hook: VerificationHook;
73
+ /** Whether the hook succeeded */
74
+ success: boolean;
75
+ /** Exit code of the command */
76
+ exitCode: number;
77
+ /** Command output (stdout + stderr) */
78
+ output: string;
79
+ /** Confidence in the result */
80
+ confidence: VerificationConfidence;
81
+ /** Execution time in milliseconds */
82
+ executionTimeMs: number;
83
+ /** Error message if hook failed to run */
84
+ error?: string;
85
+ }
86
+ /**
87
+ * Detected hooks for a project
88
+ */
89
+ export interface DetectedHooks {
90
+ /** Available hooks */
91
+ hooks: VerificationHook[];
92
+ /** Package manager detected */
93
+ packageManager: 'npm' | 'yarn' | 'pnpm' | 'unknown';
94
+ /** Whether a package.json was found */
95
+ hasPackageJson: boolean;
96
+ }
97
+ /**
98
+ * Status of a verification item
99
+ */
100
+ export type VerificationStatus = 'pending' | 'passed' | 'failed' | 'skipped' | 'not-applicable';
101
+ /**
102
+ * Confidence level in the verification result
103
+ */
104
+ export type VerificationConfidence = 'high' | 'medium' | 'low';
105
+ /**
106
+ * Method used for verification
107
+ */
108
+ export type VerificationMethod = 'automated' | 'semi-automated' | 'manual';
109
+ /**
110
+ * Result of verifying a single checklist item
111
+ */
112
+ export interface VerificationResult {
113
+ /** ID of the checklist item */
114
+ itemId: string;
115
+ /** Verification status */
116
+ status: VerificationStatus;
117
+ /** Method used for verification */
118
+ method: VerificationMethod;
119
+ /** Confidence in the result */
120
+ confidence: VerificationConfidence;
121
+ /** Evidence of verification (command output or agent reasoning) */
122
+ evidence?: string;
123
+ /** Reason for failed/skipped status */
124
+ reason?: string;
125
+ /** Timestamp of verification */
126
+ verifiedAt: string;
127
+ }
128
+ /**
129
+ * Overall status of the verification report
130
+ */
131
+ export type ReportStatus = 'pending' | 'in-progress' | 'completed' | 'requires-attention';
132
+ /**
133
+ * Summary statistics for verification
134
+ */
135
+ export interface VerificationSummary {
136
+ /** Total number of items */
137
+ total: number;
138
+ /** Number of passed items */
139
+ passed: number;
140
+ /** Number of failed items */
141
+ failed: number;
142
+ /** Number of skipped items */
143
+ skipped: number;
144
+ /** Number of not-applicable items */
145
+ notApplicable: number;
146
+ /** Coverage percentage (passed / (total - skipped - notApplicable)) */
147
+ coveragePercent: number;
148
+ /** Number of automated checks */
149
+ automatedChecks: number;
150
+ /** Number of manual checks */
151
+ manualChecks: number;
152
+ }
153
+ /**
154
+ * Full verification report
155
+ */
156
+ export interface VerificationReport {
157
+ /** Report version */
158
+ version: '1.0';
159
+ /** ID of the prompt being verified */
160
+ promptId: string;
161
+ /** Source of the prompt (fast or deep) */
162
+ source: 'fast' | 'deep';
163
+ /** When verification started */
164
+ startedAt: string;
165
+ /** When verification completed (all items done) */
166
+ completedAt?: string;
167
+ /** Overall status */
168
+ status: ReportStatus;
169
+ /** All checklist items */
170
+ items: ChecklistItem[];
171
+ /** Verification results for each item */
172
+ results: VerificationResult[];
173
+ /** Summary statistics */
174
+ summary: VerificationSummary;
175
+ /** Detected hooks used for automated verification */
176
+ detectedHooks?: DetectedHooks;
177
+ }
178
+ /**
179
+ * Intent-to-checklist mapping entry
180
+ */
181
+ export interface IntentChecklist {
182
+ /** The intent type */
183
+ intent: PromptIntent;
184
+ /** Checklist items for this intent */
185
+ items: Array<{
186
+ content: string;
187
+ group?: string;
188
+ verificationType: VerificationType;
189
+ }>;
190
+ }
191
+ /**
192
+ * Additional fields for PromptMetadata to support verification
193
+ */
194
+ export interface VerificationMetadata {
195
+ /** Whether verification is required for this prompt */
196
+ verificationRequired?: boolean;
197
+ /** Whether the prompt has been verified */
198
+ verified?: boolean;
199
+ /** Timestamp of last verification */
200
+ lastVerifiedAt?: string;
201
+ /** Path to verification report file */
202
+ verificationReportPath?: string;
203
+ }
204
+ //# sourceMappingURL=verification.d.ts.map
@@ -0,0 +1,8 @@
1
+ /**
2
+ * Clavix v4.8: Verification System Types
3
+ *
4
+ * Type definitions for the checklist verification system that ensures
5
+ * checklists generated by deep/fast modes are verified after implementation.
6
+ */
7
+ export {};
8
+ //# sourceMappingURL=verification.js.map
package/package.json CHANGED
@@ -1,6 +1,6 @@
1
1
  {
2
2
  "name": "clavix",
3
- "version": "4.7.0",
3
+ "version": "4.8.0",
4
4
  "description": "Clavix Intelligence™ for AI coding. Automatically optimizes prompts with intent detection, quality assessment, and adaptive patterns—no framework to learn. Works with Claude Code, Cursor, Windsurf, and 19+ other AI coding tools.",
5
5
  "type": "module",
6
6
  "main": "dist/index.js",