360 Secure SkillHub
360 安全技能中心

Strategic Review

核心功能：加载用户目标、记忆文档、任务看板等多源上下文，对比当日工作与既定目标，核查任务完成情况，分析重复问题/需求等规律，生成带项目标签、优先级的四类任务。典型场景：每晚11点自动运行，为次日工作规划及3点的夜间构建器提供支撑。输入为各类用户配置文档、任务数据、当日会话；输出为含进度表格、任务列表的复盘报告及MC平台任务。

h-mascot
h-mascot
热度 24.71
最近更新 2026-06-04 00:18:00
分类：数据依赖：代码执行 / 读写文件 / 第三方 API / 内部调用大模型应用场景：报告撰写 / 任务管理 / 工作流 / 数据处理 ...

安装技能

复制下面的话术发给你的 Agent，它会通过 secure-skills 检索并安装该技能。

执行命令安装此 skill： npx secure-skills-cli add ec-strategic-review
Metadata
name
strategic-review
description
End-of-day strategic review. Loads all sessions, extracts needs/opportunities, searches skills, creates MC tasks tagged by project. Feeds the overnight builder.

Strategic Review Skill

Deep strategic review that analyzes the day against Henry's goals and context.

Schedule

Runs at 11pm (before 3am overnight builder)

What It Does - DEEP ANALYSIS

Phase 1: Load Full Context

~/clawd/USER.md          → Henry's goals, preferences, current focus
~/clawd/memory.md        → Long-term facts, rules, patterns
~/clawd/memory/*.md      → Recent daily notes (last 3 days)
~/clawd/memory/proactive-work.md → Current priorities
MC board (all columns)   → Full task state
Today's sessions         → What actually happened

Phase 2: Map Today Against Goals

Questions to answer:

  1. What were Henry's stated priorities today?
  2. What did he actually work on?
  3. What's the gap?

Check against USER.md focus areas:

  • Soteria AI (primary focus) — any progress?
  • Curacel (running well) — any fires?
  • Personal — anything pending?

Phase 3: Check Completions

Scan today's requests and verify:

  • "Find flight options" → Was this done? Link to output.
  • "Fix X bug" → Was it fixed? Doc created?
  • "Build Y" → Was it shipped? Where's the PR?
  • "Research Z" → Was it completed? Where's the summary?

For each uncompleted request:

  • Create MC task with original context
  • Tag with project (Soteria/Curacel/Personal/Moltbot)

Phase 4: Pattern Analysis

Look for:

  • Recurring problems (same issue multiple times)
  • Recurring requests (same ask, different days)
  • Blockers (what kept coming up as blocked?)
  • Bugs encountered → Should there be a prevention doc?

Create improvement tasks:

  • "Document X to prevent future bugs"
  • "Automate Y since it's requested often"
  • "Build Z skill since no tool exists"

Phase 5: Goal Progress Report

For each focus area, assess:

Area Goal Today's Progress Gap Next Action
Soteria Launch GTM ? ? ?
Curacel Running smoothly ? ? ?
Personal ? ? ? ?
Moltbot $10k/mo ? ? ?

Phase 6: Create Strategic Tasks

Task types:

  1. Completion tasks — Finish uncompleted requests
  2. Prevention tasks — Docs/automation to prevent recurring issues
  3. Goal tasks — Move needle on stated priorities
  4. Gap tasks — Fill holes in coverage

Each task must have:

  • Clear link to goal/context
  • Project tag: [Soteria] [Curacel] [Personal] [Moltbot]
  • Priority: P0 (blocking) / P1 (important) / P2 (nice to have)
  • Assignee: Ada / Scotty / Spock

Usage

# Collect data
~/clawd/skills/strategic-review/scripts/collect-sessions.sh

# The cron job then:
# 1. Loads all context files
# 2. Runs the analysis
# 3. Creates MC tasks
# 4. Reports to Telegram

Output Format

🎯 Strategic Review - [Date]

## Goal Progress
[PNG TABLE: Area | Goal | Progress | Gap | Next]

## Uncompleted Requests
- ❌ Flight options (requested 2pm) → Creating task
- ✅ MC session sync (requested 12am) → Done

## Patterns Detected
- Bug: X happened 3 times → Creating prevention doc
- Request: Y asked for repeatedly → Creating automation

## Tasks Created
[PNG TABLE: # | Task | Project | Priority | Assignee]

## Tomorrow's Focus
Based on goals + gaps, tomorrow should prioritize:
1. [Most important thing]
2. [Second thing]
3. [Third thing]

Integration

  • Context: USER.md, memory.md, memory/*.md
  • Sessions: ~/.clawdbot/agents/*/sessions/
  • Skills: ~/clawd/skills/
  • MC: http://100.106.69.9:3000/api/tasks
  • Consumer: Overnight builder at 3am