convez.ai
Skill
pvieira-design

pedro-mota-skills

pvieira-design/pedro-mota-skills

Ver no GitHub
InicialPluginScore · de 0 a 100

Project & planning agent skills for Claude Code — Matt Pocock's skills (MIT) + Pedro Mota's optimized ones, with an HTML tutorial.

Instalação

$npx convez-ai pvieira-design/pedro-mota-skills

1 comando no terminal: detecta o formato do repositório (marketplace ou plugin avulso) e instala via Claude Code. Requer Node 18+ e o Claude Code instalado.

Métricas

Estrelas
1
Forks
0
Acompanhando
0
Saúde
42%
Último commit
há 3 meses
Versão
—
Issues abertas
0
Contribuidores
1

README

A tradução para o português deste README ainda está na fila.

Pedro Mota Skills

A curated bundle of project & planning agent skills for Claude Code (and compatible agents) — mixing Matt Pocock's skills with Pedro Mota's optimized ones, focused on one thing: making the agent smart about a codebase from day one through a living knowledge base and a tight plan → build → document loop.

👉 Full visual tutorial: open index.html in a browser.

The idea

The agent shouldn't re-learn the project every session. These skills install and maintain a docs/ knowledge base so it understands a feature by reading a page, doesn't reopen settled decisions, and doesn't repeat past mistakes.

ArtifactQuestion it answers
CONTEXT.mdwhat the words mean (glossary)
docs/adr/why we decided this way (decisions)
docs/system/what the code does today (living docs)
docs/plans/what we're going to do (plans)
docs/pending/what's left open (loose ends)
docs/learnings/where we already erred (lessons)

The loop

1. /grill-with-docs   decide; pin vocabulary in CONTEXT.md + write ADRs
2. /to-plan           write the plan in docs/plans/
3. /handoff           hand the plan to a NEW agent — to execute it, or to continue planning
4. implement          the new agent builds it (optionally with /tdd)
5. /sync-doc          update docs/system/ to match the shipped code
6. /to-plan done      archive the plan to docs/plans/done/

along the way:  /to-pending  (defer a loose end)   ·   docs/learnings/  (record a lesson)

/handoff is the bridge between planning and building: once the plan is written, it produces a ready-to-paste prompt (saved under .handoff/ and opened) that a new agent/chat uses to pick up — either to execute the plan or to continue the planning — wired to all the docs above.

Skills

Setup

  • setup-pedro-mota — one-shot knowledge-base bootstrap (+ calls setup-matt-pocock-skills).
  • setup-matt-pocock-skills — per-repo agent config (issue tracker, triage labels, domain layout).

Grilling & planning

  • grill-with-docs, grill-me — stress-test a plan against the domain.
  • to-plan — distill grilling decisions into a plan; close it to done/.
  • to-pending — record/resolve loose ends.
  • ubiquitous-language — extract a domain glossary.

Docs & handoff

  • sync-doc — keep docs/system/ in sync with the code.
  • handoff — emit a ready-to-paste prompt (continue planning or execute) wired to the docs.

Issues & PRDs

  • to-issues, to-prd, triage.

Engineering

  • tdd, improve-codebase-architecture, diagnose, zoom-out, prototype.

See NOTICE.md for which skills are Matt Pocock's and which are Pedro's.

Install

Skills are folders with a SKILL.md. Point your agent's skills directory at them. For Claude Code, symlink each into your skills dir:

git clone https://github.com/pvieira-design/pedro-mota-skills.git
cd pedro-mota-skills
# symlink all skills into ~/.claude/skills (adjust path to your setup)
for d in skills/*/; do ln -sfn "$PWD/$d" ~/.claude/skills/"$(basename "$d")"; done

Then invoke any skill as /<skill-name> (e.g. /setup-pedro-mota). Start a new repo with /setup-pedro-mota.

sync-doc is shipped as a skill here; in some setups it lives as a command (~/.claude/commands/sync-doc.md). Either works.

License

MIT — see LICENSE. Original skills © Matt Pocock; additions/adaptations © Pedro Mota.

Ver o original no GitHub

Bom para

Relacionados