openclaw-skills

Public repository for selected OpenClaw skill definitions developed and maintained by Amirofcodes.

Purpose

This repo showcases practical work on agent behavior design:

• task-specific skill prompts,
• operational guardrails,
• workflow structure for reliable autonomous assistance.

It is part of an active VPS-synced workflow and reflects real iteration over time.

What is in this repository

• skills/*/SKILL.md → skill-level instructions and workflow logic
• docs/openclaw/ → supporting guidance and best practices

Current public skills include:

• brainstorm
• connect-dots

Scope and boundaries

This repository is focused on skill design. It is not a full production config dump and does not aim to document every private runtime detail.

Some implementation details may be generalized to keep the repo safe for public sharing while preserving educational value.

Stability note

• Main branch updates are incremental and sync-safe.
• No history rewrites are used for normal updates.
• Breaking changes are avoided in routine edits.

Why this matters

Well-designed skills are a core part of agent reliability:

• clearer task decomposition,
• fewer hallucinated actions,
• more predictable operator outcomes.

Maintainer

Jaouad Bouddehbine (Amirofcodes)