Actionize Skill: Turn Insights Into Action

The Actionize skill (skill-actionize) turns loose insights — research findings, brainstorming output, InfraNodus gap analyses, design docs — into a concrete, scheduled plan with deadlines. It works in Claude Code and OpenClaw: you co-design the plan together, it saves everything to a .plan/ folder in your project, and sets up daily Telegram reminders that run independently of your AI agent, so nothing falls through the cracks.

It's built for anyone whose research sessions end with a great list of insights and no next steps. The skill produces plans, not code — it deliberately never implements anything itself.

Get Started

https://mcp.infranodus.com

https://mcp.infranodus.com

claude mcp add infranodus -s user --transport http https://mcp.infranodus.com/mcp

{
  "mcpServers": {
    "infranodus": {
      "url": "https://mcp.infranodus.com"
    }
  }
}

codex mcp add infranodus --url https://mcp.infranodus.com
codex mcp login infranodus

{
  "mcpServers": {
    "infranodus": {
      "command": "npx",
      "args": ["-y", "infranodus-mcp-server"],
      "env": {
        "INFRANODUS_API_KEY": "your-api-key"
      }
    }
  }
}

/plugin marketplace add infranodus/skills
/plugin install skill-actionize@infranodus-skills

npx skills add infranodus/skills --skill skill-actionize -g -a claude-code -y

npx skills add infranodus/skills --skill skill-actionize -g

What the Skill Does

Once installed, prompts like these trigger it:

• "Actionize this" or "turn this into tasks" — converts the insights from your current conversation, a file, or a design doc into a task plan.
• "Make a plan for shipping this feature by next month" — co-designs the goal, timeline, and task breakdown with you, one question at a time.
• "Schedule this" / "create deadlines" — spreads tasks across your chosen timeline with specific dates, effort estimates, and dependencies.
• "/actionize" in a project with an existing plan — shows overdue, due-today, and upcoming tasks, and lets you mark tasks done, reschedule, or add new ones.
• "/actionize diagnose" — analyzes your planning patterns across projects: what you plan, what you finish, and what you consistently avoid.

The skill also suggests itself proactively: after a research or brainstorming session that ends with insights but no concrete next steps, or when an existing plan has more than 30% deferred tasks.

How It Works

Collaborative plan design. The skill never dumps a plan on you. It asks structured questions — what's the concrete outcome, what's the deadline, does this task breakdown look right — and iterates until you approve. Each task gets a name, a "done looks like" description, a specific date, dependencies, and an effort estimate (S/M/L/XL).

Persistent .plan/ directory. The approved plan is saved in your project root: plan.md (the human-readable master plan), one markdown file per task in tasks/, and a machine-readable .status.json. A bundled done.sh CLI lets you mark tasks complete from the terminal without the agent: .plan/bin/done.sh list, .plan/bin/done.sh 3, .plan/bin/done.sh undo 3. Plans are personal — .plan/ is added to .gitignore automatically.

Telegram reminders via system cron. The skill walks you through creating a Telegram bot with @BotFather (about 2 minutes), then installs a real system crontab entry — not a session-scoped job — that sends a daily status message with overdue, due-today, and upcoming tasks. It runs whether or not your agent is open. Telegram is optional; the skill works with in-session reminders alone.

Session-start check. When you open the project again, the skill shows overdue and due-today tasks before you start working, so the plan stays in front of you.

Diagnose Mode: Analyze Your Planning Patterns

/actionize diagnose maintains a user-wide task history across all your projects and sorts every task into three states: planned, completed, and deferred (past deadline, never done). It then uses InfraNodus MCP tools to find the topical patterns in each:

Reports are saved and compared over time, so you can see whether your completion rate is improving and whether the same themes keep slipping. The output is a coach-style summary with a concrete recommendation — for example, "front-load one UI task per week: your deferrals suggest avoidance of design decisions, not lack of time." The core planning workflow works standalone; diagnose mode is where the InfraNodus MCP server comes in.

Install the Skill

Claude Code

Download the skill into your global skills folder:

cd ~/.claude/skills
curl -L -o skill-actionize.zip https://github.com/infranodus/skills/releases/latest/download/skill-actionize.zip
unzip skill-actionize.zip -d skill-actionize

Then trigger it with /actionize or just say "turn this into a plan."

Claude Web / Desktop

Go to Settings > Capabilities, enable Code execution and file creation, scroll to the Skills section, and upload the skill-actionize.zip file. Note that the cron reminders and .plan/ persistence are designed for local environments like Claude Code and OpenClaw.

OpenClaw

Run in the OpenClaw chat:

install this skill: https://github.com/infranodus/skills/releases/latest/download/skill-actionize.zip

See the main installation guide for step-by-step instructions per LLM client.

What's Next

• All Skills & Use Cases
• Skills Installation Guide
• InfraNodus Skill
• InfraNodus MCP Server