원클릭으로
structured-dev-cycle
6-stage structured development cycle with stage-based tool restrictions
Codex 또는 Claude로 설치 이 Prompt를 복사해 Codex, Claude 또는 다른 어시스턴트에 붙여 넣으면 Skill 페이지를 검토하고 설치를 진행할 수 있습니다.
메뉴
6-stage structured development cycle with stage-based tool restrictions
Codex 또는 Claude로 설치 이 Prompt를 복사해 Codex, Claude 또는 다른 어시스턴트에 붙여 넣으면 Skill 페이지를 검토하고 설치를 진행할 수 있습니다.
SOC 직업 분류 기준
Full R017 verification (5+3 rounds) before commit
Load a skill profile to switch active plugin set. Use when user wants to focus on a specific workflow (web-app/data-eng/harness-dev/minimal) and reduce skill enumeration block size per
Deploy applications to Vercel with auto-detection and preview URLs
Auto-detect project context and optimize harness — deactivate unused agents/skills, suggest missing experts, generate project profile
Monitor Claude Code releases and auto-generate GitHub issues for each new version
Parameterized evaluator-optimizer loop for quality-critical output with configurable rubrics
| name | structured-dev-cycle |
| description | 6-stage structured development cycle with stage-based tool restrictions |
| scope | core |
| version | 1.0.0 |
| user-invocable | true |
A disciplined 6-stage development cycle that enforces quality through stage-based tool restrictions. Prevents premature implementation by requiring planning and verification phases.
Inspired by Pi Coding Agent Workflow Extension's structured development approach. The core insight: restricting file modification tools during planning phases forces thorough analysis before code changes.
| # | Stage | Allowed Tools | Blocked Tools | Purpose |
|---|---|---|---|---|
| 1 | Plan | Read, Glob, Grep, WebSearch, WebFetch | Write, Edit, Bash (modifying) | Define approach, analyze requirements |
| 2 | Verify Plan | Read, Glob, Grep | Write, Edit, Bash | Review plan from different perspective |
| 3 | Implement | All tools | None | Write code, create files |
| 4 | Verify Implementation | Read, Glob, Grep, Bash (tests only) | Write, Edit | Review code, run tests |
| 5 | Compound | Read, Bash (tests only) | Write, Edit | Integration testing, cross-module validation |
| 6 | Done | Read | Write, Edit, Bash | Summary and documentation |
Following the reasoning-sandwich pattern:
| Stage | Recommended Model | Rationale |
|---|---|---|
| 1: Plan | opus | Architectural reasoning, requirement analysis |
| 2: Verify Plan | opus | Edge case detection, alternative evaluation |
| 3: Implement | sonnet | Code generation, file creation optimized |
| 4: Verify Implementation | sonnet | Test execution, structural review |
| 5: Compound | sonnet | Integration testing, cross-module validation |
| 6: Done | haiku | Checklist validation, summary generation |
Model selection is advisory — the orchestrator may override based on task complexity.
Stage state is tracked via a marker file for hook enforcement:
# Set stage (used by orchestrator or skill)
echo "plan" > /tmp/.claude-dev-stage-$PPID
# Valid stage values (all block Write/Edit except 'implement'):
# plan, verify-plan, implement, verify-impl, compound, done
# Clear stage (disable blocking)
rm -f /tmp/.claude-dev-stage-$PPID
A PreToolUse hook in .claude/hooks/hooks.json checks this marker and blocks Write/Edit tools during non-implementation stages.
[Stage 1/6: Plan]
├── Analyze requirements and constraints
├── Read existing code for context
├── Search for related patterns
├── Define approach with rationale
└── Output: Implementation plan document
Exit criteria: Clear plan with file list, approach description, and risk assessment.
[Stage 2/6: Verify Plan]
├── Review plan for completeness
├── Check for missing edge cases
├── Validate against existing patterns
├── Consider alternative approaches
└── Output: Plan approval or revision requests
Exit criteria: Plan verified by different perspective (ideally different model via multi-model-verification).
[Stage 3/6: Implement]
├── Follow verified plan
├── Create/modify files as specified
├── Write tests alongside code
├── Track deviations from plan
└── Output: Implementation complete
Exit criteria: All planned files created/modified, tests written.
[Stage 4/6: Verify Implementation]
├── Run test suite
├── Review code quality
├── Check for plan deviations
├── Validate error handling
└── Output: Verification report
Exit criteria: All tests pass, no critical issues found. If issues found, return to Stage 3.
[Stage 5/6: Compound]
├── Run integration tests
├── Cross-module validation
├── Check for side effects
├── Verify documentation accuracy
└── Output: Integration report
Exit criteria: No integration issues. If issues found, return to Stage 3.
[Stage 6/6: Done]
├── Summarize changes made
├── List files modified
├── Note any deviations from plan
├── Suggest follow-up tasks
└── Output: Completion summary
Stage 1 (Plan) maps to Claude Code's EnterPlanMode. When the structured cycle is active:
Stage 2 (Verify Plan) and Stage 4 (Verify Implementation) can invoke the multi-model-verification skill for comprehensive review.
The stage marker file (/tmp/.claude-dev-stage-$PPID) is read by a PreToolUse hook that enforces tool restrictions. This provides a safety net beyond instruction-based compliance.
For complex tasks, Agent Teams is preferred when available (R018):
When Agent Teams is enabled AND task involves 3+ agents or review→fix cycles, using Agent Teams is MANDATORY per R018.
| Task Complexity | Recommended Cycle |
|---|---|
| Simple fix (< 3 files) | Skip — direct implementation |
| Medium feature (3-10 files) | Stages 1, 3, 4, 6 (skip verify plan, compound) |
| Complex feature (10+ files) | Full 6-stage cycle |
| Architecture change | Full 6-stage cycle with multi-model verification |
| Security-critical code | Full 6-stage cycle (mandatory) |
# Orchestrator manages transitions:
echo "plan" > /tmp/.claude-dev-stage-$PPID # Enter planning
echo "verify-plan" > /tmp/.claude-dev-stage-$PPID # Enter plan verification
echo "implement" > /tmp/.claude-dev-stage-$PPID # Enter implementation
echo "verify-impl" > /tmp/.claude-dev-stage-$PPID # Enter impl verification
echo "compound" > /tmp/.claude-dev-stage-$PPID # Enter compound testing
echo "done" > /tmp/.claude-dev-stage-$PPID # Mark done
rm -f /tmp/.claude-dev-stage-$PPID # Clear (disable blocking)
/tmp/.claude-dev-stage-$PPID), so concurrent Claude Code sessions use distinct markers and do not conflict. The reader hooks (stage-blocker.sh, task-state-precompact.sh) and the writer commands above all resolve $PPID to the same Claude Code process, keeping the gate consistent./tmp/ directory is accessible to all users. The -$PPID suffix isolates per-session state but does not restrict filesystem permissions; avoid storing sensitive data in the marker file.When spawning agents via the Agent tool during this skill's execution, always pass mode: "bypassPermissions". The Agent tool default (acceptEdits) overrides agent frontmatter permissionMode, causing permission prompts during unattended execution.
═══ Structured Dev Cycle ═══════════════════════════
[■■□□□□] Stage 2/6: Verify Plan
Files planned: 5 | Risks identified: 2
═════════════════════════════════════════════════════