clavix 4.5.0 → 4.6.0

package/dist/templates/slash-commands/_canonical/archive.md

@@ -152,7 +152,7 @@ You are helping the user archive completed PRD projects to keep their workspace
 - They're still accessible in archive directory
 - Can be restored at any time
 
-## Prompts Are Separate (v2.7+)
+## Prompts Are Separate
 
 Fast/deep prompts from `/clavix:fast` and `/clavix:deep` are stored separately in `.clavix/outputs/prompts/`.
 

package/dist/templates/slash-commands/_canonical/deep.md

@@ -179,7 +179,7 @@ Deep mode provides **Clavix Intelligence™** with comprehensive analysis that g
 
 ---
 
-## Agent Transparency (v4.4)
+## Agent Transparency (v4.6)
 
 ### Quality Output Format
 {{INCLUDE:agent-protocols/quality-output.md}}
@@ -187,6 +187,9 @@ Deep mode provides **Clavix Intelligence™** with comprehensive analysis that g
 ### Agent Decision Rules
 {{INCLUDE:agent-protocols/decision-rules.md}}
 
+### Assertion Checkpoints
+{{INCLUDE:agent-protocols/assertion-checkpoints.md}}
+
 ### Patterns Applied
 {{INCLUDE:sections/pattern-visibility.md}}
 
@@ -345,7 +348,7 @@ Consider using `/clavix:prd` if this login page is part of a larger authenticati
 - **Deep mode** (`/clavix:deep`): Comprehensive analysis - best for complex prompts needing exploration
 - **PRD mode** (`/clavix:prd`): Strategic planning - best for features requiring architecture/business decisions
 
-## Next Steps (v2.7+)
+## Next Steps
 
 ### Saving the Prompt (REQUIRED)
 

package/dist/templates/slash-commands/_canonical/fast.md

@@ -158,7 +158,7 @@ Clavix provides **Clavix Intelligence™** that automatically detects intent and
 
 ---
 
-## Agent Transparency (v4.4)
+## Agent Transparency (v4.6)
 
 ### Quality Output Format
 {{INCLUDE:agent-protocols/quality-output.md}}
@@ -166,6 +166,9 @@ Clavix provides **Clavix Intelligence™** that automatically detects intent and
 ### Agent Decision Rules
 {{INCLUDE:agent-protocols/decision-rules.md}}
 
+### Assertion Checkpoints
+{{INCLUDE:agent-protocols/assertion-checkpoints.md}}
+
 ### Escalation Factors (Smart Triage)
 {{INCLUDE:sections/escalation-factors.md}}
 
@@ -208,6 +211,11 @@ Actionability: 30%
     • Too vague to start implementation
     • Missing technical specifications
 
+Specificity: 25%
+    • No file paths or identifiers
+    • Missing version numbers
+    • Uses vague term "login page" without scope
+
 Overall Quality: 42% (needs-improvement)
 
 Recommendation:
@@ -248,7 +256,7 @@ Success Criteria:
 [Actionability] Converted vague "create" into specific implementation requirements with measurable success criteria
 ```
 
-## Next Steps (v2.7+)
+## Next Steps
 
 ### Saving the Prompt (REQUIRED)
 

package/dist/templates/slash-commands/_canonical/prd.md

@@ -316,7 +316,7 @@ The validation ensures generated PRDs are immediately usable for AI consumption
 
 ---
 
-## Agent Transparency (v4.4)
+## Agent Transparency (v4.6)
 
 ### Quality Output Format
 {{INCLUDE:agent-protocols/quality-output.md}}
@@ -330,6 +330,9 @@ The validation ensures generated PRDs are immediately usable for AI consumption
 ### Agent Decision Rules
 {{INCLUDE:agent-protocols/decision-rules.md}}
 
+### Assertion Checkpoints
+{{INCLUDE:agent-protocols/assertion-checkpoints.md}}
+
 ---
 
 ## Troubleshooting

package/dist/templates/slash-commands/_components/agent-protocols/assertion-checkpoints.md

@@ -0,0 +1,122 @@
+## Assertion Checkpoints (v4.6)
+
+At key workflow stages, verify your execution matches expected state. These checkpoints help ensure correct template execution.
+
+### Fast/Deep Mode Checkpoints
+
+**CHECKPOINT 1: After Intent Detection**
+```
+✓ Intent type identified (one of 11 types)
+✓ Confidence percentage calculated
+✓ If confidence < 70%: Secondary intent noted
+```
+
+**CHECKPOINT 2: After Quality Assessment**
+```
+✓ All 6 dimensions scored:
+  - Clarity (0-100%)
+  - Efficiency (0-100%)
+  - Structure (0-100%)
+  - Completeness (0-100%)
+  - Actionability (0-100%)
+  - Specificity (0-100%)
+✓ Overall quality calculated
+✓ Strengths identified (any dimension >= 85%)
+```
+
+**CHECKPOINT 3: After Optimization**
+```
+✓ Enhanced prompt generated
+✓ Improvements listed with dimension labels
+✓ Patterns applied documented
+```
+
+**CHECKPOINT 4: After Saving**
+```
+✓ Prompt saved to .clavix/outputs/prompts/{mode}/
+✓ Index file updated
+✓ Success message displayed
+```
+
+### PRD Mode Checkpoints
+
+**CHECKPOINT 1: Before Questions**
+```
+✓ State assertion displayed
+✓ Implementation blocked message shown
+```
+
+**CHECKPOINT 2: After Each Question**
+```
+✓ Answer validated for completeness
+✓ Follow-up asked if answer is vague
+✓ Minimum requirements met before proceeding
+```
+
+**CHECKPOINT 3: After Document Generation**
+```
+✓ Full PRD generated with all sections
+✓ Quick PRD generated (2-3 paragraphs)
+✓ Quality validation scores displayed
+```
+
+**CHECKPOINT 4: After Saving**
+```
+✓ Files saved to .clavix/outputs/{project-name}/
+✓ full-prd.md exists
+✓ quick-prd.md exists
+✓ Next steps displayed
+```
+
+### Implementation Mode Checkpoints
+
+**CHECKPOINT 1: Before Starting**
+```
+✓ Config file checked (.clavix-implement-config.json)
+✓ Resume state detected if exists
+✓ Tasks loaded from tasks.md
+```
+
+**CHECKPOINT 2: After Each Task**
+```
+✓ Task completed successfully
+✓ clavix task-complete {task-id} called
+✓ Next task displayed automatically
+```
+
+**CHECKPOINT 3: After All Tasks**
+```
+✓ All tasks marked complete
+✓ Archive suggestion displayed
+✓ Git commit created if configured
+```
+
+### Verification Block Format
+
+At the end of your response, include this verification block:
+
+```
+## Clavix Execution Verification
+- [x] Intent detected: {type} ({confidence}%)
+- [x] Quality assessed: {overall}%
+- [x] {N} patterns applied
+- [x] Prompt saved: {filename}
+- [x] Mode: {fast|deep|prd|plan}
+```
+
+This allows users and developers to verify correct template execution and helps identify any deviations from expected behavior.
+
+### Self-Verification Protocol
+
+If a checkpoint fails:
+1. **STOP**: Do not proceed past the failed checkpoint
+2. **IDENTIFY**: Which checkpoint failed and why
+3. **REPORT**: Inform user of the issue
+4. **RECOVER**: Take corrective action or ask for guidance
+
+Example recovery:
+```
+⚠️ Checkpoint 2 failed: Quality assessment incomplete
+Issue: Specificity dimension not scored
+Action: Rerunning quality assessment with all 6 dimensions
+```

package/dist/templates/slash-commands/_components/agent-protocols/decision-rules.md

@@ -184,6 +184,34 @@ IF user prompt is empty/invalid:
   → NEVER: Proceed with assumption
 ```
 
+### Rule 11: Execution Verification (v4.6)
+
+```
+BEFORE completing response:
+  → INCLUDE verification block at end
+  → VERIFY all checkpoints met for current mode
+
+  IF any checkpoint failed:
+    → REPORT which checkpoint failed
+    → EXPLAIN why it failed
+    → SUGGEST recovery action
+
+  IF all checkpoints passed:
+    → SHOW verification block with all items checked
+```
+
+**Verification Block Template:**
+```
+## Clavix Execution Verification
+- [x] Intent detected: {type} ({confidence}%)
+- [x] Quality assessed: {overall}%
+- [x] {N} patterns applied
+- [x] Prompt saved: {filename}
+- [x] Mode: {fast|deep|prd|plan}
+```
+
+---
+
 ### Rule Summary Table
 
 | Condition | Action | User Communication |
@@ -202,3 +230,4 @@ IF user prompt is empty/invalid:
 | 3+ strategic keywords | Suggest PRD | "Strategic scope detected" |
 | pattern fails | Skip + report | "Pattern [X] skipped" |
 | file write fails | Retry then report | "Error: [details]" |
+| response complete | Include verification | Show checkpoint status |

package/dist/templates/slash-commands/_components/sections/pattern-visibility.md

@@ -137,7 +137,9 @@ v4.3.2 Conversational mode (deep):
 
 | Mode | Patterns Available | Typical Applied |
 |------|-------------------|-----------------|
-| Fast | 12 core patterns | 4-7 patterns |
-| Deep | 27 total patterns | 8-14 patterns |
-| PRD | 15 patterns | 10-15 patterns |
-| Conversational | 11 patterns | 6-11 patterns |
+| Fast | 12 patterns | 4-7 patterns |
+| Deep | 27 patterns | 8-14 patterns |
+| PRD | 12 patterns | 8-12 patterns |
+| Conversational | 3 patterns | 1-3 patterns |
+
+**Note:** Deep mode has access to all 27 patterns. Pattern selection varies by intent - some patterns only apply to specific intents (e.g., PRDStructureEnforcer only for prd-generation), so typical applied count is 8-14.

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "clavix",
-  "version": "4.5.0",
+  "version": "4.6.0",
   "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.",
   "type": "module",
   "main": "dist/index.js",