메뉴
Use when executing implementation plans with independent tasks in the current session using subagents
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
| name | ito-subagent-driven-development |
| description | Use when executing implementation plans with independent tasks in the current session using subagents |
Execute plan by dispatching fresh subagent per task, with two-stage review after each: spec compliance review first, then code quality review.
Core principle: Fresh subagent per task + two-stage review (spec then quality) = high quality, fast iteration
ito-apply)vs. ito-apply:
ito tasks complete <change-id> <task-id>ito-finish# Get the change context
ito agent instruction apply --change <change-id>
# Read tasks.md
ITO_ROOT="$(ito path ito-root)"
cat "$ITO_ROOT/changes/<change-id>/tasks.md"
# Extract all tasks with full text and context upfront
ito tasks start <change-id> <task-id>
Provide:
Pick an implementer agent tier based on task complexity:
ito-quick: small, localized changesito-general: most tasks (default)ito-thinking: complex refactors, tricky bugs, high-risk editsImplementer uses TDD:
Dispatch spec reviewer subagent with:
Reviewer checks:
If issues: implementer subagent fixes, re-review until ✅
Dispatch code quality reviewer subagent with:
Reviewer checks:
If issues: implementer subagent fixes, re-review until ✅
ito tasks complete <change-id> <task-id>
If more tasks: repeat from step 1
If done: dispatch final reviewer, then use ito-finish
./implementer-prompt.md - Dispatch implementer subagent./spec-reviewer-prompt.md - Dispatch spec compliance reviewer subagent./code-quality-reviewer-prompt.md - Dispatch code quality reviewer subagentNever:
If subagent asks questions:
If reviewer finds issues:
Required workflow skills:
ito-using-git-worktrees - Set up isolated workspace before startingito-proposal - Creates the plan this skill executesito-requesting-code-review - Code review template for reviewer subagentsito-finish - Complete development after all tasksSubagents should use:
ito-test-driven-development - Subagents follow TDD for each taskAlternative workflow:
ito-apply - Use for human-in-loop execution with batch checkpointsManage remote Kocao agent sessions from an AI assistant. Use when: listing running agents, starting a new remote agent, checking agent status, viewing agent logs, stopping an agent session, or sending a prompt to a running remote agent.
Fast trigram-indexed code search across the entire repository using Zoekt (the engine behind GitHub and Sourcegraph code search). Use INSTEAD OF grep/ripgrep for broad codebase searches. Supports regex, literal strings, file filters, and symbol queries. Auto-installs zoekt binaries and auto-indexes the repo on first use.
Apply a Change Proposal. Triggered by the user saying "Apply change <change-id>" or "Implement change <change-id>". Use when implementing, executing, applying, building, coding, or developing a feature, change, requirement, enhancement, fix, or modification. Use when running tasks from a spec, proposal, or plan.
You MUST use this before any creative work - creating features, proposing changes,building components, adding functionality, or modifying behavior. Explores user intent, requirements and design before implementation.
Use when implementation is complete, all tests pass, and you need to decide how to integrate the work — presents structured options for merge, PR, or cleanup
Run an ito ralph loop for a specific change (or module/repo sequence), with safe defaults and automatic restart context on early exits.