claude-fsd 1.5.25 → 1.5.26

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.
package/README.md CHANGED
@@ -69,10 +69,6 @@ claude-fsd create-plan # Generate development plan from requirements
69
69
  claude-fsd --working-dir=project1 dev
70
70
  claude-fsd --working-dir=backend interview
71
71
 
72
- # Fast mode (skips slow o3-pro planning)
73
- claude-fsd create-plan --fast
74
- claude-fsd --working-dir=project1 --fast create-plan
75
-
76
72
  # claudefsd also works the same way
77
73
  claudefsd # Same as claude-fsd
78
74
  ```
@@ -87,7 +83,7 @@ Runs the development agent fleet. This command:
87
83
  - Updates the plan to track progress
88
84
  - Repeats until all tasks are done
89
85
 
90
- Every 4th cycle, it activates "megathinking mode" using the Opus model for architectural planning.
86
+ Every 4th cycle, it activates "megathinking mode" using ultrathink for deep architectural planning.
91
87
 
92
88
  #### claudefsd-interview
93
89
  Interactive expert Q&A session that:
@@ -101,8 +97,7 @@ Generates a comprehensive development plan based on:
101
97
  - Your project brief (BRIEF.md)
102
98
  - Interview answers (QUESTIONS.md) or requirements (REQUIREMENTS.md)
103
99
  - Creates PLAN.md with prioritized tasks and CLAUDE-NOTES.md with architectural analysis
104
- - Uses Opus model for standard planning, o3-pro for deep strategic analysis (when codex available)
105
- - Use `--fast` flag to skip o3-pro step for quicker results
100
+ - Uses Claude Sonnet with ultrathink for deep strategic planning
106
101
 
107
102
  ## How it Works
108
103
 
@@ -188,21 +183,18 @@ your-project/
188
183
 
189
184
  ## Model Selection Strategy
190
185
 
191
- The system intelligently selects AI models based on task complexity:
186
+ The system uses Claude Sonnet throughout with strategic ultrathinking for complex work:
192
187
 
193
- - **o3-pro Model**: Used for the most complex strategic work (requires codex CLI)
194
- - Deep architectural planning (create-plan, unless --fast used)
195
- - Takes 5-15 minutes but provides exceptional strategic reasoning
196
-
197
- - **Opus Model**: Used for complex architectural work
198
- - Requirements gathering (interview)
199
- - Standard architecture planning (create-plan with --fast or when codex unavailable)
188
+ - **Sonnet + Ultrathink**: Used for complex architectural work requiring deep thinking
189
+ - Requirements consolidation (interview)
190
+ - Deep strategic planning (create-plan)
200
191
  - Megathinking mode (every 4th dev iteration)
201
-
202
- - **Sonnet Model**: Used for regular development
192
+
193
+ - **Sonnet (Standard)**: Used for regular development
203
194
  - Standard development tasks
204
195
  - Code implementation and reviews
205
196
  - Most development iterations
197
+ - Question generation in interview
206
198
 
207
199
 
208
200
  ## License
package/bin/claudefsd CHANGED
@@ -10,7 +10,6 @@ NC='\033[0m' # No Color
10
10
 
11
11
  # Parse command line parameters
12
12
  WORKING_DIR="docs"
13
- FAST_MODE=""
14
13
  for arg in "$@"; do
15
14
  case $arg in
16
15
  --working-dir=*)
@@ -18,10 +17,6 @@ for arg in "$@"; do
18
17
  # Remove this argument from the list
19
18
  shift
20
19
  ;;
21
- --fast)
22
- FAST_MODE="--fast"
23
- shift
24
- ;;
25
20
  esac
26
21
  done
27
22
  export CLAUDEFSD_WORKING_DIR="$WORKING_DIR"
@@ -271,7 +266,7 @@ if [ $# -eq 0 ]; then
271
266
  if [ -f "$WORKING_DIR/REQUIREMENTS.md" ] || [ -f "$WORKING_DIR/QUESTIONS.md" ]; then
272
267
  echo -e "${GREEN}Creating plan from project inputs...${NC}"
273
268
  echo
274
- exec "$(dirname "$0")/claudefsd-create-plan" --working-dir="$WORKING_DIR" $FAST_MODE
269
+ exec "$(dirname "$0")/claudefsd-create-plan" --working-dir="$WORKING_DIR"
275
270
  else
276
271
  echo -e "${RED}No requirements or questions found. Please run the interview first.${NC}"
277
272
  exit 1
@@ -48,7 +48,7 @@ Read all of these documents if they exist:
48
48
  - docs/WEBTESTS.md -- the project web tests
49
49
  - README.md -- the project README
50
50
 
51
- Your job, as a megathinker business analyst, is to analyze the project brief and generate clarifying questions.
51
+ Your job, as an ultrathinker business analyst, is to analyze the project brief and generate clarifying questions.
52
52
 
53
53
  1. Read through the BRIEF.md and any existing documents to understand the project.
54
54
  2. Generate 10-15 relevant questions to clarify ambiguous aspects of the brief.
@@ -67,8 +67,8 @@ DO NOT answer the questions yourself - just generate them for the user to answer
67
67
  "
68
68
 
69
69
  # run BA's
70
- echo "Running claude with opus model..."
71
- claude --model opus --dangerously-skip-permissions -p "$prompt1" | tee >(cat > $LOGFILE-ba1)
70
+ echo "Running claude with sonnet model with ultrathink..."
71
+ claude --model sonnet --dangerously-skip-permissions -p "$prompt1" | tee >(cat > $LOGFILE-ba1)
72
72
 
73
73
  # Only run codex if available
74
74
  if command -v codex >/dev/null 2>&1; then
@@ -13,26 +13,13 @@ command_exists() {
13
13
 
14
14
  # Function to check dependencies
15
15
  check_dependencies() {
16
- local check_codex=${1:-true} # Check codex by default
17
16
  local missing_deps=()
18
- local warnings=()
19
-
20
- # Check for claude
17
+
18
+ # Check for claude (only required dependency)
21
19
  if ! command_exists claude; then
22
20
  missing_deps+=("claude")
23
21
  fi
24
-
25
- # Check for codex (optional but recommended)
26
- if [ "$check_codex" = true ] && ! command_exists codex; then
27
- warnings+=("codex is not installed. Some review features will be unavailable.")
28
- fi
29
-
30
- # Check for OPENAI_API_KEY (optional for codex)
31
- if [ "$check_codex" = true ] && [ -z "${OPENAI_API_KEY:-}" ]; then
32
- warnings+=("OPENAI_API_KEY is not set. Codex features will be limited.")
33
- warnings+=("Set it with: export OPENAI_API_KEY='your-key-here'")
34
- fi
35
-
22
+
36
23
  # Show critical errors
37
24
  if [ ${#missing_deps[@]} -ne 0 ]; then
38
25
  echo -e "${RED}❌ Missing required dependencies:${NC}"
@@ -45,15 +32,6 @@ check_dependencies() {
45
32
  echo
46
33
  exit 1
47
34
  fi
48
-
49
- # Show warnings
50
- if [ ${#warnings[@]} -ne 0 ]; then
51
- echo -e "${YELLOW}⚠️ Optional dependencies:${NC}"
52
- for warning in "${warnings[@]}"; do
53
- echo -e " - $warning"
54
- done
55
- echo
56
- fi
57
35
  }
58
36
 
59
37
  # If sourced directly, run the check
@@ -4,17 +4,12 @@ set -e
4
4
 
5
5
  # Parse command line parameters
6
6
  WORKING_DIR="docs"
7
- FAST_MODE=false
8
7
  for arg in "$@"; do
9
8
  case $arg in
10
9
  --working-dir=*)
11
10
  WORKING_DIR="${arg#*=}"
12
11
  shift
13
12
  ;;
14
- --fast)
15
- FAST_MODE=true
16
- shift
17
- ;;
18
13
  esac
19
14
  done
20
15
  export CLAUDEFSD_WORKING_DIR="$WORKING_DIR"
@@ -92,81 +87,24 @@ INFRASTRUCTURE PROPORTIONALITY RULES:
92
87
  - Choose infrastructure complexity appropriate to solution size
93
88
  "
94
89
 
95
- # run BA's
96
- echo "Running claude with opus model..."
97
- claude --model opus --dangerously-skip-permissions -p "$prompt2" | tee >(cat > $LOGFILE-ba3)
98
-
99
- # Only run codex if available and not in fast mode
100
- if command -v codex >/dev/null 2>&1 && [ "$FAST_MODE" = false ]; then
101
- echo ""
102
- echo -e "\033[33m==================================================================\033[0m"
103
- echo -e "\033[33m== RUNNING O3-PRO FOR DEEP STRATEGIC PLANNING\033[0m"
104
- echo -e "\033[33m==================================================================\033[0m"
105
- echo "⏳ This will take several minutes as o3-pro deeply analyzes the project..."
106
- echo " o3-pro excels at strategic thinking and will create a comprehensive plan."
107
- echo ""
108
-
109
- # Prepare comprehensive context for o3-pro with file references
110
- file_list=""
111
- [ -f "$BRIEF_FILE" ] && file_list="$file_list\n- $BRIEF_FILE (project brief)"
112
- [ -f "$WORKING_DIR/QUESTIONS.md" ] && file_list="$file_list\n- $WORKING_DIR/QUESTIONS.md (interview questions & answers)"
113
- [ -f "$WORKING_DIR/REQUIREMENTS.md" ] && file_list="$file_list\n- $WORKING_DIR/REQUIREMENTS.md (consolidated requirements)"
114
- [ -f "$WORKING_DIR/INTERVIEW-SESSION.json" ] && file_list="$file_list\n- $WORKING_DIR/INTERVIEW-SESSION.json (interview metadata)"
115
- [ -f "$WORKING_DIR/CLAUDE-NOTES.md" ] && file_list="$file_list\n- $WORKING_DIR/CLAUDE-NOTES.md (existing technical notes)"
116
- [ -f "$WORKING_DIR/IDEAS.md" ] && file_list="$file_list\n- $WORKING_DIR/IDEAS.md (ideas backlog)"
117
- [ -f "$WORKING_DIR/README.md" ] && file_list="$file_list\n- $WORKING_DIR/README.md (project readme)"
118
-
119
- o3_prompt="You are an elite software architect and strategic planner using o3-pro's advanced reasoning capabilities.
120
-
121
- READ AND ANALYZE THESE PROJECT FILES:$file_list
122
-
123
- TASK:
124
- Using your deep strategic reasoning, create a comprehensive project architecture and implementation plan.
125
-
126
- 1. ARCHITECTURAL ANALYSIS ($WORKING_DIR/CLAUDE-NOTES.md):
127
- - System architecture and key design decisions
128
- - Technology stack justification
129
- - Critical technical challenges and mitigation strategies
130
- - Integration points and dependencies
131
- - Performance and scalability considerations
132
- - Security architecture
133
- - Long-term maintainability strategy
134
-
135
- 2. IMPLEMENTATION PLAN ($WORKING_DIR/PLAN.md):
136
- - Master plan limited to 100 lines maximum with high-level sections
137
- - For complex projects, create detailed sub-plans in separate files ($WORKING_DIR/plan-section1.md, etc.)
138
- - Phased development approach with clear milestones
139
- - Task breakdown with dependencies (detailed in sub-plans if needed)
140
- - Risk assessment for each phase
141
- - Testing strategy integrated into each phase
142
- - Infrastructure needs (proportional to project size)
143
- - Performance benchmarks and acceptance criteria
144
-
145
- Remember:
146
- - Think strategically about the entire system lifecycle
147
- - Consider edge cases and failure modes
148
- - Plan for iterative development and feedback loops
149
- - Keep infrastructure proportional to project complexity
150
- - Prioritize robustness and maintainability
151
-
152
- Take your time to think deeply about the optimal approach."
153
-
154
- echo "Running codex o3-pro (this may take 3-5 minutes)..."
155
- codex -m o3-pro --full-auto -q "$o3_prompt" | tee $LOGFILE-ba4
156
- echo ""
157
- echo "✅ o3-pro strategic planning complete!"
158
- else
159
- echo ""
160
- if [ "$FAST_MODE" = true ]; then
161
- echo -e "\033[33m⚡ Fast mode enabled - skipping o3-pro strategic analysis for quicker results.\033[0m"
162
- echo "Use 'claudefsd-create-plan' without --fast for deep o3-pro planning."
163
- echo "Fast mode: skipping o3-pro strategic analysis" > $LOGFILE-ba4
164
- else
165
- echo -e "\033[33mNote: Install 'codex' CLI to enable o3-pro strategic planning for enhanced results.\033[0m"
166
- echo "Proceeding with standard planning using Claude Opus..."
167
- echo "Codex not available, skipping o3-pro strategic analysis" > $LOGFILE-ba4
168
- fi
169
- fi
90
+ # Add ultrathink instruction to the beginning
91
+ ultrathink_instruction="<ultrathink>
92
+ You are entering deep strategic planning mode. Before responding, engage in extended reasoning to:
93
+ - Analyze the full scope and implications of the project
94
+ - Consider architectural tradeoffs and long-term consequences
95
+ - Identify edge cases, failure modes, and technical risks
96
+ - Think through the optimal approach from multiple angles
97
+ Take your time to think deeply about the best solution before generating your plan.
98
+ </ultrathink>
99
+
100
+ "
101
+
102
+ # Prepend ultrathink to the prompt
103
+ prompt2="$ultrathink_instruction$prompt2"
104
+
105
+ # Run planning with sonnet + ultrathink
106
+ echo "Running claude with sonnet model (ultrathink mode)..."
107
+ claude --model sonnet --dangerously-skip-permissions -p "$prompt2" | tee >(cat > $LOGFILE-ba3)
170
108
 
171
109
  echo -e "\033[32m==================================================================\033[0m"
172
110
  echo -e "\033[32m== PLAN CREATION COMPLETE\033[0m"
package/bin/claudefsd-dev CHANGED
@@ -94,8 +94,20 @@ while true; do
94
94
  if [ $((LOOP_COUNTER % 4)) -eq 0 ]; then
95
95
  echo -e "\033[33m**** MEGATHINKING MODE ACTIVATED ****\033[0m"
96
96
  echo -e "\033[33mThis is your 4th development cycle. Taking a step back for architectural planning.\033[0m"
97
- MEGATHINKING_MODE="**** MEGATHINKING MODE ACTIVATED ****\nThis is your 4th development cycle. Before proceeding with the next task, please take a step back and use megathinking mode to architecturally plan the next phase of development. Consider the overall structure of the codebase, potential refactoring opportunities, design patterns, technical debt, and how the current work connects to broader project goals.\n\n"
98
- CLAUDE_MODEL="opus"
97
+ MEGATHINKING_MODE="<ultrathink>
98
+ **** MEGATHINKING MODE ACTIVATED ****
99
+ This is your 4th development cycle. Before proceeding with the next task, engage in deep extended reasoning to:
100
+ - Take a step back and architecturally plan the next phase of development
101
+ - Consider the overall structure of the codebase
102
+ - Identify potential refactoring opportunities
103
+ - Evaluate design patterns and technical debt
104
+ - Think about how the current work connects to broader project goals
105
+ - Analyze edge cases and potential failure modes
106
+ Take your time to think deeply about the optimal approach before proceeding.
107
+ </ultrathink>
108
+
109
+ "
110
+ CLAUDE_MODEL="sonnet"
99
111
  else
100
112
  MEGATHINKING_MODE=""
101
113
  CLAUDE_MODEL="sonnet"
@@ -384,7 +384,7 @@ Create a comprehensive requirements document that:
384
384
  Format the output as a clean, well-structured markdown document suitable for feeding into project planning.
385
385
  "
386
386
 
387
- claude --model opus -p "$consolidation_prompt" > "$REQUIREMENTS_FILE"
387
+ claude --model sonnet -p "$consolidation_prompt" > "$REQUIREMENTS_FILE"
388
388
 
389
389
  echo -e "${GREEN}Requirements consolidated in $REQUIREMENTS_FILE${NC}"
390
390
 
package/package.json CHANGED
@@ -1,6 +1,6 @@
1
1
  {
2
2
  "name": "claude-fsd",
3
- "version": "1.5.25",
3
+ "version": "1.5.26",
4
4
  "description": "Claude Full Self Drive tools for autonomous AI-powered development",
5
5
  "bin": {
6
6
  "claude-fsd": "bin/claude-fsd",
@@ -1,69 +0,0 @@
1
- #!/bin/bash
2
- #
3
- # Main development dispatcher - now uses unified development mode by default
4
- # Usage: claudefsd-dev [direct|iterative|legacy-direct|legacy-iterative]
5
- #
6
- # Modes:
7
- # (default): Unified mode combining best of both approaches
8
- # direct: Legacy single-context parallel execution (deprecated)
9
- # iterative: Legacy multi-iteration loop development (deprecated)
10
- # legacy-direct: Force legacy direct execution mode
11
- # legacy-iterative: Force legacy iterative loop mode
12
- #
13
- # Examples:
14
- # claudefsd-dev # Use unified development mode (recommended)
15
- # claudefsd-dev legacy-direct # Force legacy direct execution mode
16
- # claudefsd-dev legacy-iterative # Force legacy iterative loop mode
17
-
18
- set -e
19
-
20
- # Parse command line arguments
21
- DEV_MODE="$1"
22
-
23
- # Default to unified mode
24
- if [ -z "$DEV_MODE" ]; then
25
- DEV_MODE="unified"
26
- fi
27
-
28
- echo "Development mode: $DEV_MODE"
29
-
30
- # Route to appropriate development script
31
- case "$DEV_MODE" in
32
- unified)
33
- echo "Launching unified development mode (recommended)..."
34
- exec "$(dirname "$0")/claudefsd-dev-unified"
35
- ;;
36
- direct)
37
- echo "WARNING: Legacy direct mode is deprecated. Consider using unified mode."
38
- echo "Launching legacy direct execution mode..."
39
- exec "$(dirname "$0")/claudefsd-dev-direct"
40
- ;;
41
- iterative)
42
- echo "WARNING: Legacy iterative mode is deprecated. Consider using unified mode."
43
- echo "Launching legacy iterative development mode..."
44
- exec "$(dirname "$0")/claudefsd-dev-iterative"
45
- ;;
46
- legacy-direct)
47
- echo "Launching legacy direct execution mode..."
48
- exec "$(dirname "$0")/claudefsd-dev-direct"
49
- ;;
50
- legacy-iterative)
51
- echo "Launching legacy iterative development mode..."
52
- exec "$(dirname "$0")/claudefsd-dev-iterative"
53
- ;;
54
- small)
55
- # Legacy support - map to unified (was direct)
56
- echo "Legacy 'small' mode - redirecting to unified development..."
57
- exec "$(dirname "$0")/claudefsd-dev-unified"
58
- ;;
59
- large)
60
- # Legacy support - map to unified (was iterative)
61
- echo "Legacy 'large' mode - redirecting to unified development..."
62
- exec "$(dirname "$0")/claudefsd-dev-unified"
63
- ;;
64
- *)
65
- echo "Unknown development mode: $DEV_MODE"
66
- echo "Valid modes: unified (default), direct, iterative, legacy-direct, legacy-iterative"
67
- exit 1
68
- ;;
69
- esac