Execute qualquer Skill no Manus
com um clique

Execute qualquer Skill no Manus com um clique

create-skill

Estrelas6

Forks0

Atualizado8 de fevereiro de 2026 às 02:50

Create a new Claude Code skill from a description. Use when the user says "create skill", "new skill", or "add a slash command".

Instalação

Instalar com Codex ou Claude Copie este prompt, cole no Codex, Claude ou outro assistente e deixe que ele revise a página da skill e instale para você.

Executar no Manus

Fonte

opmau

opmau/claude-code-framework

Abrir repositório GitHub Ver repositórios do creator

Download

Executar no Manus

Ocupações relacionadasSOC

Baseado na classificação ocupacional SOC

Desenvolvedores de softwareInformática e Matemática·SOC 15-1252

SKILL.md

readonly

name	create-skill
description	Create a new Claude Code skill from a description. Use when the user says "create skill", "new skill", or "add a slash command".
argument-hint	<skill-name>
user-invocable	true
allowed-tools	Read, Write, Glob
model	sonnet

/create-skill — Generate a new skill

Create a new skill (slash command) following the framework's conventions.

Steps

If $ARGUMENTS is empty, ask the user for:
- Skill name (lowercase, hyphenated)
- What the skill should do (one sentence)
- When it should trigger (what the user would say)
Read an existing skill for reference patterns:
```
.claude/skills/review/SKILL.md
```

Generate the skill file following this structure:

---
name: <skill-name>
description: <what it does>. Use when the user says "<trigger phrases>".
argument-hint: "<hint>"
user-invocable: true
allowed-tools: <minimal set needed>
model: <haiku for simple, sonnet for complex>
---

Body structure:

# /<skill-name> — <one-line description>

<What this skill does, in one sentence.>

## Steps

1. <first action>
2. <second action>
3. <report results in specific format>

## Arguments

- No arguments: <default behavior>
- `$ARGUMENTS`: <behavior with arguments>

## Notes

- <important constraint 1>
- <important constraint 2>

Write the skill to .claude/skills/<skill-name>/SKILL.md

Report what was created:

Created: .claude/skills/<skill-name>/SKILL.md
Trigger: /<skill-name>
Description: <what it does>

Guidelines

Context efficiency: Keep skills under 100 lines. The context window is a shared resource.
Model selection: Use haiku for simple tasks (build, test, file checks). Use sonnet for tasks requiring analysis (review, planning).
Tool minimization: Only include tools the skill actually needs in allowed-tools.
No side effects: Skills should report findings — not silently make changes unless that's their explicit purpose.
Follow conventions: Match the tone and format of existing skills in the project.

Mais deste repositório

mesmo repositório

fix-issue

opmau/claude-code-framework

Pick a bug from Linear, fix it test-first, verify, and update. Use when the user says "fix issue", "fix bug", "work on known issue", or "pick a bug".

2026-04-146

create-ticket

opmau/claude-code-framework

Create a new ticket for task tracking. Use when the user says "create ticket", "new ticket", or "track this".

2026-04-096

smoke-test

opmau/claude-code-framework

Diagnose an integration/E2E test log, classify failures by severity, and batch-create issues. Use when the user says "smoke test", "diagnose log", "triage test log", or provides a test log.

2026-04-096

linear-sync

opmau/claude-code-framework

Pull Linear issues and generate a local snapshot for quick reference. Use when the user says "sync linear", "pull linear issues", "update from linear", or "show my issues".

2026-02-096

linear-triage

opmau/claude-code-framework

Triage, prioritize, and groom Linear issues. Use when the user says "triage linear", "review inbox", "prioritize issues", "groom backlog", or "clean up issues".

2026-02-096

diagnose

opmau/claude-code-framework

Investigate a bug using structured differential diagnosis. Use when the user says "diagnose", "investigate this", "what's causing this", or "figure out why".

2026-02-096

name	create-skill
description	Create a new Claude Code skill from a description. Use when the user says "create skill", "new skill", or "add a slash command".
argument-hint	<skill-name>
user-invocable	true
allowed-tools	Read, Write, Glob
model	sonnet

/create-skill — Generate a new skill

Create a new skill (slash command) following the framework's conventions.

Steps

If $ARGUMENTS is empty, ask the user for:
- Skill name (lowercase, hyphenated)
- What the skill should do (one sentence)
- When it should trigger (what the user would say)
Read an existing skill for reference patterns:
```
.claude/skills/review/SKILL.md
```

Generate the skill file following this structure:

---
name: <skill-name>
description: <what it does>. Use when the user says "<trigger phrases>".
argument-hint: "<hint>"
user-invocable: true
allowed-tools: <minimal set needed>
model: <haiku for simple, sonnet for complex>
---

Body structure:

# /<skill-name> — <one-line description>

<What this skill does, in one sentence.>

## Steps

1. <first action>
2. <second action>
3. <report results in specific format>

## Arguments

- No arguments: <default behavior>
- `$ARGUMENTS`: <behavior with arguments>

## Notes

- <important constraint 1>
- <important constraint 2>

Write the skill to .claude/skills/<skill-name>/SKILL.md

Report what was created:

Created: .claude/skills/<skill-name>/SKILL.md
Trigger: /<skill-name>
Description: <what it does>

Guidelines

Context efficiency: Keep skills under 100 lines. The context window is a shared resource.
Model selection: Use haiku for simple tasks (build, test, file checks). Use sonnet for tasks requiring analysis (review, planning).
Tool minimization: Only include tools the skill actually needs in allowed-tools.
No side effects: Skills should report findings — not silently make changes unless that's their explicit purpose.
Follow conventions: Match the tone and format of existing skills in the project.