Execute qualquer Skill no Manus
com um clique

Execute qualquer Skill no Manus com um clique

crafting-skills

Estrelas17

Forks2

Atualizado25 de julho de 2026 às 12:16

Use when creating or updating skills in `~/.config/ai/skills`, especially when a `SKILL.md` has a weak trigger, tangled structure, poor steering, or bloated wording.

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

l-lin

l-lin/dotfiles

Abrir repositório GitHub Ver repositórios do creator

Download

Executar no Manus

Ocupações relacionadasSOC

Baseado na classificação ocupacional SOC

Outras ocupações de informáticaInformática e Matemática·SOC 15-1299

Explorador de arquivos

3 arquivos

SKILL.md

readonly

name	crafting-skills
description	Use when creating or updating skills in `~/.config/ai/skills`, especially when a `SKILL.md` has a weak trigger, tangled structure, poor steering, or bloated wording.
disable-model-invocation	true

REQUIRED SUB-SKILL: Use clear-writing for the final rewrite.

Read the target SKILL.md and any linked references. Do not edit yet.
Interview the user using ask-user-question tool before changing anything. Cover all four: trigger, structure, steering, pruning. Use references/SKILL-CHECKLIST.md.
Decide the smallest honest move: create, rewrite, split, merge, or delete. Ask before split, merge, or delete.
Rewrite for the chosen path.
- Trigger: choose user-invoked vs model-invocable on purpose. For user-only skills, set disable-model-invocation: true.
- Structure: keep SKILL.md tiny. Use numbered steps. Move branch-only reference into supporting files.
- Steering: pick a few leading words and repeat them consistently.
- Pruning: delete sediment, duplication, stale guidance, and no-ops.
Bias toward rewrite-after-audit when the old skill is bloated or tangled. Patch in place only when the structure is already sound.
Hand off verification. Name the pressure scenarios that should fail without the skill, the deletion tests that prove each paragraph earns its keep, and the next check to run. Do not claim the skill works without evidence.

Mais deste repositório

mesmo repositório

kibana-elasticsearch

l-lin/dotfiles

Use when querying Elasticsearch through the Kibana console proxy, especially for log investigations, field-based filtering, or error triage from job or application logs.

2026-07-2517

devils-advocate

l-lin/dotfiles

Use when pressure-testing a plan, architecture, code change, or AI-generated proposal before committing, especially when hidden assumptions, production risks, or failure modes may be underexplored.

2026-07-2517

datadog-pup

l-lin/dotfiles

Use when interacting with Datadog resources using the `pup` CLI, especially for APM traces, metrics, monitors, notebooks, and synthetic tests.

2026-07-2517

code-simplifier

l-lin/dotfiles

Use when simplifying recently touched code without changing behavior, especially when asked to "simplify code", "clean up code", "refactor for clarity", or "improve readability".

2026-07-2517

ingest-url

l-lin/dotfiles

Ingest a URL into the patient account wiki. Use when the user provides a URL (Atlassian/Confluence, web articles, documentation) and wants to add it to their research wiki. Handles fetching, markdown conversion, asset extraction, and runs the full wiki ingest workflow. Triggers on "ingest <url>", "add this to the wiki", or when user shares a URL in the context of wiki research.

2026-07-2417

journal

l-lin/dotfiles

Use at session start, after context compaction, and before session end. Owns session journals in `.sandbox/journals/`: when to create one, what prior context to read, what to log, and how to leave a resumable handoff

2026-07-1117

name	crafting-skills
description	Use when creating or updating skills in `~/.config/ai/skills`, especially when a `SKILL.md` has a weak trigger, tangled structure, poor steering, or bloated wording.
disable-model-invocation	true

REQUIRED SUB-SKILL: Use clear-writing for the final rewrite.

Read the target SKILL.md and any linked references. Do not edit yet.
Interview the user using ask-user-question tool before changing anything. Cover all four: trigger, structure, steering, pruning. Use references/SKILL-CHECKLIST.md.
Decide the smallest honest move: create, rewrite, split, merge, or delete. Ask before split, merge, or delete.
Rewrite for the chosen path.
- Trigger: choose user-invoked vs model-invocable on purpose. For user-only skills, set disable-model-invocation: true.
- Structure: keep SKILL.md tiny. Use numbered steps. Move branch-only reference into supporting files.
- Steering: pick a few leading words and repeat them consistently.
- Pruning: delete sediment, duplication, stale guidance, and no-ops.
Bias toward rewrite-after-audit when the old skill is bloated or tangled. Patch in place only when the structure is already sound.
Hand off verification. Name the pressure scenarios that should fail without the skill, the deletion tests that prove each paragraph earns its keep, and the next check to run. Do not claim the skill works without evidence.