claude-flow-novice 2.18.10 → 2.18.12

package/.claude/skills/cfn-cerebras-code-generator/SKILL.md

@@ -1,13 +1,27 @@
+---
+name: cfn-cerebras-code-generator
+description: "FAST code generation via Z.ai glm-4.6 model. Use for rapid test generation, boilerplate code, repetitive patterns, and bulk file creation. Ideal when speed matters more than nuance. Do NOT use for complex architectural decisions or security-critical code."
+version: 2.0.0
+tags: [code-generation, fast, zai, glm-4.6, tests, boilerplate]
+---
+
 # Cerebras Code Generator Skill
 
 ## Description
-Generates code using Cerebras API with context tracking and history management. Acts as an OpenAI-compatible wrapper for Cerebras API.
+Generates code using Z.ai glm-4.6 model for **fast test and code generation**. Use this for rapid iteration when generating tests, boilerplate, and repetitive code patterns.
+
+## When to Use
+- ✅ **Test generation** - unit tests, integration tests, test fixtures
+- ✅ **Boilerplate code** - CRUD operations, API endpoints, data models
+- ✅ **Repetitive patterns** - similar components, migration scripts
+- ✅ **Bulk file creation** - generating multiple similar files quickly
+- ❌ **NOT for** complex architecture, security-critical code, or nuanced logic
 
 ## Configuration
 ```bash
 # Required environment variables
-export CEREBRAS_API_KEY="your-api-key"
-export CEREBRAS_MODEL="qwen2.5-coder-32b"  # or other Cerebras model
+export ZAI_API_KEY="your-api-key"  # or CEREBRAS_API_KEY for legacy
+export ZAI_MODEL="glm-4.6"  # Fast, cost-effective model
 
 # Optional settings
 export CEREBRAS_BASE_URL="https://api.cerebras.ai/v1"

package/.claude/skills/cfn-cerebras-coordinator/SKILL.md

@@ -1,7 +1,14 @@
+---
+name: cfn-cerebras-coordinator
+description: "Coordinates FAST code generation via Z.ai glm-4.6 with RuVector pattern learning. Use when agents need rapid test generation, bulk code creation, or repetitive boilerplate. Tracks successful prompts for continuous improvement. Ideal for high-volume, low-complexity code tasks."
+version: 2.0.0
+tags: [coordination, code-generation, fast, zai, glm-4.6, pattern-learning, tests]
+---
+
 # Cerebras Coordinator Skill
 
 ## Description
-Integrates Cerebras code generation with RuVector for prompt tracking, success pattern learning, and agent coordination. Agents use this skill to offload code generation to Cerebras while maintaining a searchable database of what works.
+Coordinates **fast code generation** via Z.ai glm-4.6 model with RuVector pattern learning. Agents use this skill to offload rapid test generation and boilerplate code while building a searchable database of successful patterns.
 
 ## Key Features
 - 🚀 **Fast Code Generation**: Uses Cerebras API for rapid code creation
@@ -65,15 +72,22 @@ Agent (Coordinator)        Cerebras Coordinator Skill         RuVector
         |                           |                              |
 ```
 
+## When to Use
+- ✅ **Bulk test generation** - generating many test files quickly
+- ✅ **Boilerplate with patterns** - learning from previous successful generations
+- ✅ **Agent code offloading** - when agents need fast, simple code generation
+- ✅ **Repetitive tasks** - migrations, similar components, data models
+- ❌ **NOT for** complex logic, security code, or architectural decisions
+
 ## Configuration
 
 ```bash
 # Required
-export CEREBRAS_API_KEY="your-cerebras-api-key"
+export ZAI_API_KEY="your-zai-api-key"  # or CEREBRAS_API_KEY for legacy
 export RUVECTOR_INDEX_PATH="./.claude/skills/cfn-ruvector-codebase-index/data"
 
 # Optional
-export CEREBRAS_MODEL="qwen2.5-coder-32b"
+export ZAI_MODEL="glm-4.6"  # Fast, cost-effective model (default)
 export COORDINATION_DB_PATH="./.claude/skills/cfn-cerebras-coordinator/generations.db"
 export DEFAULT_TEST_TIMEOUT="60"
 export MAX_GENERATION_ATTEMPTS="3"

package/.claude/skills/cfn-cerebras-mcp/SKILL.md

@@ -1,10 +1,24 @@
+---
+name: cfn-cerebras-mcp
+description: "FAST code generation via mcp__cerebras-mcp__write tool using Z.ai glm-4.6. Use for rapid test generation, boilerplate creation, and bulk code tasks in main chat. Prompt must be SHORTER than output. Ideal for tests, CRUD, migrations, and repetitive patterns."
+version: 2.0.0
+tags: [mcp, code-generation, fast, zai, glm-4.6, tests, main-chat]
+---
+
 # Cerebras MCP Code Generation
 
-Fast code generation via `mcp__cerebras-mcp__write` tool.
+**FAST** code generation via `mcp__cerebras-mcp__write` tool using Z.ai glm-4.6 model.
 
 ## When to Use
 
-Use for generating or modifying code files when speed matters. Prompt must be SHORTER than expected output.
+Use for **rapid test and code generation** when speed matters more than nuance:
+- ✅ **Test files** - unit tests, integration tests, test fixtures
+- ✅ **Boilerplate** - CRUD endpoints, data models, components
+- ✅ **Bulk creation** - multiple similar files quickly
+- ✅ **Migrations** - database migrations, schema updates
+- ❌ **NOT for** complex architecture, security code, nuanced logic
+
+**Rule**: Prompt must be SHORTER than expected output (blueprint style).
 
 ## Usage
 

package/.claude/skills/cfn-compilation-error-fixer/package.json

@@ -25,11 +25,9 @@
     "prepare": "chmod +x bin/fix-errors.sh",
     "postinstall": "echo 'CFN Compilation Error Fixer installed successfully! Run: npm run fix:rust or npm run fix:ts'"
   },
-  "dependencies": {},
   "optionalDependencies": {
-    "@cerebras/cerebras_cloud_sdk": "^1.0.0"
+    "@cerebras/cerebras_cloud_sdk": "^1.59.0"
   },
-  "devDependencies": {},
   "engines": {
     "node": ">=18.0.0"
   },
@@ -44,4 +42,4 @@
   ],
   "author": "CFN Team",
   "license": "MIT"
-}
+}

package/.claude/skills/cfn-local-ruvector-accelerator/index-code.sh

@@ -89,6 +89,9 @@ fi
 # Convert to absolute path
 PATH_TO_INDEX="$(cd "$PATH_TO_INDEX" && pwd)"
 
+# Convert array to comma-separated string for Python
+FILE_TYPES_STR=$(IFS=','; echo "${FILE_TYPES[*]}")
+
 echo "🔍 Indexing code patterns in: $PATH_TO_INDEX"
 echo "📄 File types: ${FILE_TYPES[*]}"
 if [[ -n "$PATTERNS" ]]; then
@@ -113,7 +116,7 @@ def generate_pattern_id(file_path, content):
 def extract_functions(content, file_type):
     \"\"\"Extract functions/classes from code\"\"\"
     patterns = []
-    
+
     # Simple extraction based on language
     if file_type == 'rs':
         # Rust functions, structs, impls
@@ -121,10 +124,10 @@ def extract_functions(content, file_type):
         fn_pattern = r'^(pub\s+)?(async\s+)?(unsafe\s+)?fn\s+(\w+)'
         struct_pattern = r'^(pub\s+)?struct\s+(\w+)'
         impl_pattern = r'^impl\s+(\w+)\s*(for\s+(\w+))?'
-        
+
         for i, line in enumerate(content.split('\\n')):
-            for pattern, type_name in [(fn_pattern, 'function'), 
-                                      (struct_pattern, 'struct'), 
+            for pattern, type_name in [(fn_pattern, 'function'),
+                                      (struct_pattern, 'struct'),
                                       (impl_pattern, 'impl')]:
                 match = re.match(pattern, line)
                 if match:
@@ -134,13 +137,13 @@ def extract_functions(content, file_type):
                         'line': i + 1,
                         'signature': line.strip()
                     })
-    
+
     elif file_type == 'py':
         # Python functions, classes
         import re
         fn_pattern = r'^\s*def\s+(\w+)'
         class_pattern = r'^\s*class\s+(\w+)'
-        
+
         for i, line in enumerate(content.split('\\n')):
             if re.match(fn_pattern, line):
                 patterns.append({
@@ -151,41 +154,53 @@ def extract_functions(content, file_type):
                 })
             elif re.match(class_pattern, line):
                 patterns.append({
-                    'type': 'class', 
+                    'type': 'class',
                     'name': re.match(class_pattern, line).group(1),
                     'line': i + 1,
                     'signature': line.strip()
                 })
-    
+
     return patterns
 
 # Initialize engine
 engine = SearchEngine('${STORAGE_PATH}/storage')
 
+# Parse file types from shell
+file_types = '${FILE_TYPES_STR}'.split(',')
+verbose = '${VERBOSE}' == 'true'
+
 # Index files
 indexed_count = 0
 pattern_count = 0
 
 print('📂 Scanning files...')
-for file_type in ${FILE_TYPES[@]}:
+for file_type in file_types:
+    file_type = file_type.strip()
+    if not file_type:
+        continue
     pattern = f'**/*.{file_type}'
     files = list(Path('${PATH_TO_INDEX}').rglob(pattern))
-    
+
     for file_path in files:
+        # Skip node_modules, .git, dist, build directories
+        path_str = str(file_path)
+        if any(skip in path_str for skip in ['node_modules', '.git', '/dist/', '/build/', '/target/']):
+            continue
+
         if file_path.is_file():
             try:
                 with open(file_path, 'r', encoding='utf-8') as f:
                     content = f.read()
-                
+
                 if len(content) < 50:  # Skip very small files
                     continue
-                
+
                 # Generate pattern ID
                 pattern_id = generate_pattern_id(str(file_path), content)
-                
+
                 # Extract functions/classes
                 patterns = extract_functions(content, file_type)
-                
+
                 # Prepare metadata
                 metadata = {
                     'file_size': len(content),
@@ -193,7 +208,7 @@ for file_type in ${FILE_TYPES[@]}:
                     'patterns': patterns,
                     'last_modified': os.path.getmtime(file_path)
                 }
-                
+
                 # Store pattern
                 engine.add_pattern(
                     pattern_id=pattern_id,
@@ -202,15 +217,15 @@ for file_type in ${FILE_TYPES[@]}:
                     content=content,
                     metadata=metadata
                 )
-                
+
                 indexed_count += 1
                 pattern_count += len(patterns)
-                
-                if ${VERBOSE}:
+
+                if verbose:
                     print(f'  ✅ {file_path} ({len(patterns)} patterns)')
-                    
+
             except Exception as e:
-                if ${VERBOSE}:
+                if verbose:
                     print(f'  ⚠️  Skipped {file_path}: {e}')
                 continue
 

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "claude-flow-novice",
-  "version": "2.18.10",
+  "version": "2.18.12",
   "description": "Claude Flow Novice - Advanced orchestration platform for multi-agent AI workflows with CFN Loop architecture\n\nIncludes Local RuVector Accelerator and all CFN skills for complete functionality.",
   "main": "index.js",
   "type": "module",

package/scripts/postinstall.js

@@ -138,7 +138,9 @@ function main() {
 }
 
 // Run the script
-main().catch(error => {
+try {
+  main();
+} catch (error) {
   console.error('CFN post-install failed:', error);
   process.exit(1);
-});
+}