菜单
Legacy compatibility alias for task-router-lite. Keeps older codex-max references working while using a thin Phase PLAN routing model.
用 Codex 或 Claude 帮你安装 复制这段 Prompt，粘贴到 Codex、Claude 或其他助手里，让它检查 Skill 页面并帮你完成安装。
基于 SOC 职业分类
MANDATORY: Run appropriate validation tools after EVERY code change. Do not finish a task until the code is error-free.
Implements Manus-style file-based planning to organize and track progress on complex tasks. Creates task_plan.md, findings.md, and progress.md. Use when asked to plan out, break down, or organize a multi-step project, research task, or any work requiring 5+ tool calls. Supports automatic session recovery after /clear.
Use when encountering any bug, test failure, or unexpected behavior, before proposing fixes
Thin Phase PLAN router for codex-max. Select at most one primary execution skill and one optional secondary skill from the installed local skill surface.
Use when a significant technical decision should be recorded with context, alternatives, and consequences so the reasoning survives beyond the current session.
Use before creative or constructive work when requirements or solution shape are still unclear. Turns vague ideas into confirmed designs before implementation begins.
| name | task-intelligence |
| description | Legacy compatibility alias for task-router-lite. Keeps older codex-max references working while using a thin Phase PLAN routing model. |
| risk | safe |
| source | codex-max |
| date_added | 2026-04-19 |
This skill name is kept only for backward compatibility. It no longer behaves like a broad multi-agent amplification or orchestration layer.
Current behavior: act as the thin Phase PLAN router used by task-router-lite.
task-router-lite as the canonical PLAN behavior.task-intelligence working without changing the runtime contract abruptly.0-1 primary execution skill0-1 optional secondary skill| Situation | Primary skill |
|---|---|
| Bug, failing test, unexpected behavior | systematic-debugging |
| Multi-step or high-context task | planning-with-files |
| Medium planning task | concise-planning |
| Validation after code changes | lint-and-validate |
| Completion / fixed / passed claim | verification-before-completion |
| Underspecified feature | brainstorming if installed; otherwise concise-planning or planning-with-files |
| Library / framework / API lookup | documentation-lookup if installed; otherwise local docs / repo evidence / Context7 |
| Before coding, dependency, or utility choice | search-first if installed; otherwise targeted repository search |
| Unfamiliar repo onboarding | codebase-onboarding if installed; otherwise targeted repository discovery |
| Significant architecture decision or trade-off | architecture-decision-records if installed; otherwise ADR-style structured notes |
| Git branching / commit / PR / merge / release discipline | git-workflow if installed; otherwise repo git conventions |
| Critical user journey or browser-flow verification | e2e-testing if installed; otherwise existing test tools |
| Auth / secrets / sensitive endpoints / risky data handling | security-review if installed; otherwise direct security checklist |
Use a secondary skill only when it clearly adds value and keep it to one:
verification-before-completion for completion gateslint-and-validate after code changesplanning-with-files for persistent state on long tasksBefore implementation, produce the workspace readiness table:
| Context Element | Details |
|---|---|
| Active repo | [path] |
| Related repos | [paths] |
| Context readiness | Ready / Incomplete |
| Execution intent | [brief description] |
| Difficulty level | easy, medium, or hard |
| Subagent decision | Spawning Subagent / Direct Execution |
| Phase BUILD Skills | [selected skills] |
agent-orchestrator scripts.When all workspace references have migrated, task-intelligence should remain only as a compatibility alias or be retired cleanly.