Skip to content
Back to skills

Loop Create

ASecurity

Scaffold a new agent with the full file structure. Use when creating any new agent for the Energy platform. NOTE — ".loop" is a working name for the agent file format. When building, use a configurable extension constant.

  • 9 stars
  • 0 votes
  • 0 copies
  • 0 views
  • Added October 20, 2026
ai-agents

Security analysis

A100/100

Scanned October 20, 2026

npx -y skills add naman10parikh/agentgrid --skill loop-create --agent claude-code

Installs into .claude/skills of the current project.

Are you the author of Loop Create?

Add the live security badge to your README. It updates with every re-scan.

Security grade badge for Loop Create
[![Security: A — Skills Directory](https://www.skillsdirectory.com/api/skills/naman10parikh-loop-create/badge)](https://www.skillsdirectory.com/skills/naman10parikh-loop-create)

More formats (shields.io, HTML) on the badges page. Keep it an A: scan every change in CI with Pro.

Download with Pro
SKILL.md
---
name: loop-create
description: Scaffold a new agent with the full file structure. Use when creating any new agent for the Energy platform. NOTE — ".loop" is a working name for the agent file format. When building, use a configurable extension constant.
---

## Required Input

Ask the user for:

1. **Agent name** — symbolic, one word (e.g., Ribbon, Compass, Quill, Forge)
2. **Workflow** — one sentence describing what it automates
3. **UI type** — chat | dashboard | tool-picker | flow-builder | game | custom
4. **Target user** — who uses this? (e.g., "couples planning date nights")

## Generated Structure

```
{AgentName}.loop/
├── SOUL.md           # Identity, personality, boundaries
├── skills/
│   └── core.md       # Primary workflow with steps
├── MEMORY.md         # Bootstrap memory (<2000 tokens)
├── HEARTBEAT.md      # Periodic checks and self-maintenance
├── ui-template.yaml  # UI type, color scheme, layout
└── BRAND.md          # Name rationale, tagline, landing copy
```

## SOUL.md Template

```markdown
# {AgentName}

## Identity

I am {AgentName}, powered by Energy. I help {target_user} by {workflow}.

## Personality

- Tone: [warm/professional/playful/direct]
- I use [first person / "we"]
- I proactively suggest next steps
- I remember past interactions

## Boundaries

- I never share user data across sessions without consent
- I escalate to a human when confidence < 70%
- I refuse harmful or illegal requests
```

## Naming Convention

The name should be a symbolic element of the workflow:

- Ribbon (gifts) — wrapping, presentation
- Compass (dates) — direction, exploration
- Quill (content) — writing instrument
- Forge (dev) — creation, building

After generating, verify all files are parseable and memory is under 2,000 tokens.

Attribution

Is this your skill, or is something wrong with this listing? Request removal or report an issue. Author removals are honored within 72 hours.

Comments

Loading comments…

Newsletter

Get the weekly list of safe new skills, and the ones we pulled.