ClawKit Logo
ClawKitReliability Toolkit
Back to Registry
Official Verified

agentic-loop-designer

Skill by flynndavid

skill-install — Terminal

Install via CLI (Recommended)

clawhub install openclaw/skills/skills/flynndavid/agentic-loop-designer
Or

Agentic Loop Designer

Framework: The Yes/No Loop Canvas Worth $300/hr consultant time. Yours for $29.


What This Skill Does

Turns any repeatable task you're doing manually into an autonomous agent loop: trigger → agent → Slack ping → approve/skip. Includes 5 ready-to-deploy loop templates and a decision tree to design your own from scratch.

Problem it solves: Founders spend hours on task sequences that could run themselves. Email triage, weekly reports, standup summaries, lead qualification — if you do it the same way more than twice a week, it should be a loop.


The Yes/No Loop Canvas

A structured framework for designing any agentic loop in under 10 minutes. Built around one insight: every automatable task is just a series of yes/no decisions.

Canvas Structure

┌─────────────────────────────────────────────┐
│              YES/NO LOOP CANVAS             │
├─────────────────────────────────────────────┤
│  TRIGGER: What starts this loop?           │
│  ─────────────────────────────────────────  │
│  AGENT ACTION: What does the agent do?     │
│  ─────────────────────────────────────────  │
│  DECISION GATE: Approve / Skip / Escalate? │
│  ─────────────────────────────────────────  │
│  OUTPUT: What gets created or sent?        │
│  ─────────────────────────────────────────  │
│  MEMORY: What should persist for next run? │
└─────────────────────────────────────────────┘

Step 1: Trigger Qualification

What kicks off this loop?

Is there a natural trigger?
├── Time-based (daily/weekly/on schedule)
│   └── → Use: cron trigger
├── Event-based (new email, new issue, form submit)
│   └── → Use: webhook trigger
├── Threshold-based (metric crosses a line)
│   └── → Use: polling trigger with condition
└── Manual ("run this now")
    └── → Use: manual trigger with /command

Trigger scoring:

Trigger TypeReliabilitySetup EffortRecommended For
Cron★★★★★LowReports, summaries, digests
Webhook★★★★☆MediumNew data events
Polling★★★☆☆MediumMetric-based
Manual★★★★★NoneOn-demand workflows

Step 2: Agent Action Design

What should the agent actually do?

The action block answers three questions:

  1. What data does the agent need? (sources)
  2. What transformation happens? (the work)
  3. What's the output format? (how it lands)

Action Template:

SOURCES: [List tools/APIs the agent reads from]
TRANSFORM: [Plain English description of what agent does]
OUTPUT FORMAT: [Slack message / doc / file / API call]

Example:

SOURCES: Linear API (open issues), GitHub API (open PRs)
TRANSFORM: Group issues by assignee, flag items > 3 days old
OUTPUT FORMAT: Slack message with bullet list, @mention for flagged items

Step 3: Decision Gate Design

The most important part of any agentic loop.

Metadata

Stars2387
Views0
Updated2026-03-09
View Author Profile
AI Skill Finder

Not sure this is the right skill?

Describe what you want to build — we'll match you to the best skill from 16,000+ options.

Find the right skill
Add to Configuration

Paste this into your clawhub.json to enable this plugin.

{
  "plugins": {
    "official-flynndavid-agentic-loop-designer": {
      "enabled": true,
      "auto_update": true
    }
  }
}
Safety NoteClawKit audits metadata but not runtime behavior. Use with caution.