一键导入
decide
Decision support — criteria, scenarios, patterns, recommendation.
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。
菜单
Decision support — criteria, scenarios, patterns, recommendation.
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
End a session — extract learnings, write handoff, store patterns, detect doc drift. Use when ending a work session or when the user says done/bye/closing.
Build anything — features, bug fixes, refactors, spikes, migrations. Auto-detects strategy, integrates with backlog, enforces TDD. The unified dev command.
Align a project to brana practices — assess gaps, plan, implement, verify. Auto-detects type. Use when setting up a new project or realigning an existing one.
Scan and diagnose a project, or scaffold a new client from scratch. Works for code and venture clients. Auto-detects project type.
Manage the backlog — plan, track, navigate phases and streams. Use when planning phases, viewing roadmaps, or restructuring work.
Structured bug fix — reproduce (failing test), diagnose, fix (minimal change), verify, commit. Enforces test-first. Use when a bug needs a methodical fix.
| name | decide |
| description | Decision support — criteria, scenarios, patterns, recommendation. |
| effort | low |
| keywords | ["decision","recommendation","what-to-do","prioritize","choose","next","options","trade-off"] |
| task_strategies | ["spike","investigation"] |
| stream_affinity | ["roadmap","research"] |
| argument-hint | [question or options, e.g. 'should I do A or B' / 'what to work on next'] |
| group | thinking |
| model | sonnet |
| allowed-tools | ["Bash","Read","Glob","Grep","AskUserQuestion","mcp__ruflo__memory_search_unified","mcp__ruflo__autopilot_predict","ToolSearch"] |
| status | stable |
| growth_stage | evergreen |
One sentence in, one recommendation out.
Use the skill arg if provided. If the conversation makes the question obvious, use that (Recommended). Otherwise ask:
AskUserQuestion: "What's the decision? One sentence."
ToolSearch("select:mcp__ruflo__memory_search_unified")
Before answering, classify the question:
Simple (fast lane → skip to Step 4):
Complex (ask first):
For complex questions, surface the right framework before answering:
AskUserQuestion:
question: "This has multiple alternatives/dimensions. How deep do you want to go?"
header: "Depth"
options:
- label: "Quick take — best guess now"
description: "I give a recommendation with brief reasoning. Fast, opinionated."
- label: "Decision matrix — weighted scoring"
description: "We score each option against criteria with weights. Explicit, defensible. Best for vendor/tool/strategy selection."
- label: "SWOT first — strategic framing"
description: "Analyze strengths/weaknesses/opportunities/threats before deciding. Best for product or business direction."
- label: "Six hats — full perspective sweep"
description: "Run White/Yellow/Black/Green/Red/Blue lenses. Best for decisions that need both optimism and caution."
If user picks quick take → proceed to Step 4.
If user picks decision matrix → invoke decision-matrix skill via Skill tool.
If user picks SWOT → invoke swot-analysis skill via Skill tool.
If user picks six hats → invoke six-thinking-hats skill via Skill tool.
If task selection ("what to work on?"): run brana backlog next 2>/dev/null | head -5.
If ruflo is available: mcp__ruflo__memory_search_unified(query: "{QUESTION}", namespace: "pattern", limit: 2, threshold: 0.3).
Skip if neither adds value for the question.
**{question restated in ≤10 words}**
Do {X}. {One sentence why.}
No headers. No bullet lists. No "it depends" without an immediate answer. If uncertain, say "Not enough signal — try X first."