review
Skill0 starsChanged 27 days agoMITView on GitHub
Fluxo Review etapa 1 — gera revisao detalhada de um PR aberto, usando agentes especializados por stack e foco. Comentarios inline propostos para validacao do usuario.
192 other projects use a version of this.
Adopted by
Most recently taken up by
- microsoft/veronatheir own version, first seen Nov 8, 2026
- microsoft/FluidFrameworktheir own version, first seen Nov 8, 2026
- charliee1w/consolidation-memorytheir own version, first seen Nov 8, 2026
- jisuzzzz/haechitheir own version, first seen Nov 8, 2026
- apokamo/kajitheir own version, first seen Nov 8, 2026
- dirien/yet-another-agent-harnesstheir own version, first seen Nov 8, 2026
Used with
Projects with it also use close-spec, design-with-images and 10 more.
Other files people use alongside this one.
Files found in the same public projects as a version of this one, more often than they turn up anywhere else. Each project counts once, large collections count for less, and this file's own repository is left out.
dzhngSkillclose-spec
When a spec under specs/<feature>/ is done shipping (a write-spec build, or any planned task that produced a spec), archive it to specs/done/ and rewrite it from a build-plan into a durable rationale — the why, the principles, the invariants — pointing back to the real code for the how. Use when implementation has landed and the plan no longer matches what shipped, or the user says a feature/spec is finished. Pairs with [write-spec](../write-spec/SKILL.md) (the plan this closes) and [review](../review/SKILL.md) (run before closing).
Also in 5 projects that use this
dzhngSkilldesign-with-images
Explore UI and visual design with image generation, then iterate the real implementation against the user-selected concept. Use when brainstorming visual directions, offering mockup variants, or implementing an approved generated design; do not stop at the mockups or the first code approximation.
Also in 5 projects that use this
LambenthanSkilldaily-arxiv
每日从 arXiv 拉取新论文,过滤相关性,auto-ingest 高优先级论文,检测 SOTA 更新
Also in 5 projects that use this
LambenthanSkillexp-eval
实验判决门:Review LLM 独立评判实验结果 → 4 种判决路径 → 自动更新 claims confidence、ideas status、graph edges
Also in 5 projects that use this
- +3 more from dzhng
- +5 more from Lambenthan
Other versions192 projects use a version of this.
Word for word · 0
None found.
Edited · 0
None found.
Wrote their own version · 192
- garrytan/gstack★ 141ksame name
- ansible/ansible★ 71ksame name
- docling-project/docling★ 70ksame name
- traefik/traefik★ 65ksame name
- ClickHouse/ClickHouse★ 51ksame name
- temporalio/temporal★ 24ksame name
- tisfeng/Easydict★ 15ksame name
- juicedata/juicefs★ 15ksame name
- elie222/inbox-zero★ 13ksame name
- gitkraken/vscode-gitlens★ 10ksame name
182 more not listed
How matches are found. Word for word: the exact same text. Edited: shares passages with this file (“99% same” means 99% of its five-word phrases also appear here; 10% or more counts). Their own version: similar meaning, or, weakest, the same name. We can’t tell which project had it first. Projects are counted among the 200 most-starred candidates, and mentions are links to the repository we found, up to 50 per site, so both counts are a floor.
In this repository28 other files it gives its agents.
- ado-closeSkill
- ado-commentSkill
- ado-getSkill
- ado-refineSkill
- adrSkill
- approve-reviewSkill
- appSkill
- autoSkill
- challenge-specSkill
- completeSkill
- dart-clean-archSkill
- discoverySkill
- fixSkill
- flutter-clean-archSkill
- implementSkill
- jira-refineSkill
- kickoffSkill
- linear-refineSkill
- orcamentoSkill
- planSkill
- post-reviewSkill
- pr-openSkill
- pr-statusSkill
- refineSkill
- specifySkill
- statusSkill
- tasksSkill
- verifySkill
The file
We split this skill into the rules it gives an agent
1 rule. 0 are common, 0 are distinctive, and 1 is very specific to this project.
0
Common
Most of the 157 edited versions we read keep these.
0
Distinctive
This author's own choices. Worth a look.
1
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Very specific: line 114
Use exatamente esse numero (c + offset) no campo line.
“Common” is measured: kept by at least 6 in 10 of the 157 edited versions of this file we could read. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
1. Detectar projeto
PROJECT_PATH=$(git rev-parse --show-toplevel 2>/dev/null || pwd)
source ~/.claude/bin/get-project.sh
PROJECT_NAME=$(get-project-name)
REVIEWS_DIR="$HOME/.claude/workflow/$PROJECT_NAME/reviews"
mkdir -p "$REVIEWS_DIR"Confirme que gh esta autenticado:
gh auth statusSe nao estiver, peca: "Execute gh auth login antes de continuar."
2. Determinar PR alvo
Se o usuario passou um numero de PR como argumento, use esse numero diretamente.
Se nao passou, liste PRs abertos:
gh pr list --json number,title,author,headRefName,updatedAt --limit 30 \
--jq '.[] | "#\(.number) — \(.title) (@\(.author.login), \(.headRefName))"'Apresente a lista numerada e pergunte qual PR revisar. Aguarde a resposta.
Guarde como PR_NUM.
3. Coletar contexto do PR
gh pr view "$PR_NUM" --json number,title,body,baseRefName,headRefName,author,files \
> "$REVIEWS_DIR/PR-${PR_NUM}.meta.json"
gh pr diff "$PR_NUM" > "$REVIEWS_DIR/PR-${PR_NUM}.diff"Leia ambos os arquivos. Liste os arquivos modificados:
jq -r '.files[].path' "$REVIEWS_DIR/PR-${PR_NUM}.meta.json"4. Detectar stacks afetadas
Classifique cada arquivo modificado em uma ou mais stacks pelas regras abaixo. Um arquivo pode pertencer a mais de uma stack (ex: .dart em flutter package → flutter E dart).
| Padrao | Stack |
*.dart em arvore com pubspec.yaml que depende de flutter | flutter |
*.dart (todos) | dart |
*.ts ou *.js em arvore com package.json contendo @nestjs/* | nestjs |
android/**, *.kt, *.java (fora de teste flutter) | android |
ios/**, *.swift, *.m, *.mm, *.h (fora de flutter pods) | ios |
*.tf, cloudbuild*.yaml, app.yaml, .gcloudignore, arquivos referenciando gcloud/firebase | gcp |
*.py | python |
*.go (usa focos correctness, concurrency, security, testing) | go |
Se nenhum padrao bater para um arquivo (ex: docs, config), ignore-o do review tecnico.
Stacks unicas detectadas → DETECTED_STACKS.
5. Verificar agentes disponiveis — GATE
Para cada stack detectada, verifique se existem os 5 agentes de foco:
FOCUSES=(correctness architecture performance security testing)
MISSING=()
for stack in "${DETECTED_STACKS[@]}"; do
for focus in "${FOCUSES[@]}"; do
test -f ~/.claude/agents/${stack}-${focus}.md || MISSING+=("${stack}-${focus}")
done
doneSe MISSING estiver vazio: prossiga.
Se faltar algum agente: PAUSE o fluxo e avise:
"Detectei a(s) stack(s)
<lista>neste PR, mas faltam os seguintes agentes:
<stack>-<focus>- ...
Quero criar esses agentes com voce antes de continuar a revisao. Para cada agente faltante, preciso entender:
- Quais sao os principais riscos/padroes desse foco nessa stack neste seu contexto?
- Existem regras especificas (lint, convencoes, anti-patterns) que devem ser sempre flagged?
Posso comecar pelo
<primeiro-da-lista>?"
Aguarde a definicao do usuario. Crie cada agente em ~/.claude/agents/<nome>.md antes de prosseguir. So volte ao passo 6 quando todos os agentes existirem.
6. Executar revisao com agentes
Para cada combinacao (stack, focus) em DETECTED_STACKS x FOCUSES:
- Filtre os arquivos modificados que pertencem a essa stack
- Lance o agente
<stack>-<focus>em paralelo (mesma message com variosAgentcalls quando possivel) - Passe ao agente:
- Lista de arquivos relevantes (paths absolutos)
- Caminho do diff:
$REVIEWS_DIR/PR-${PR_NUM}.diff - Caminho do mapa da org (
<org>/.map/index.md), se existir no repo ou numa pasta acima: o agente usa para contratos cross-repo, rotas, providers e endpoints em vez de grep amplo Instrucao: "Leia o diff e os arquivos. Retorne JSON array de comentarios inline. Schema:
[{path, line, severity, body}]. severity ∈ critical|major|minor|nit. Sem prosa fora do JSON.IMPORTANTE sobre
line: o numero deve ser a linha no arquivo NEW (pos-imagem), nao o offset dentro do diff. Para encontrar a linha correta:- Procure o hunk header
@@ -a,b +c,d @@que contem o trecho que voce quer comentar. - A primeira linha de conteudo apos esse header corresponde a linha
cdo arquivo NEW. - Cada linha que comeca com
+ou(espaco/contexto) avanca o contador da linha NEW; linhas que comecam com-nao avancam. - Use exatamente esse numero (
c + offset) no campoline. - Para arquivos novos (
@@ -0,0 +1,N @@), a linha NEW vai de 1 ate N. - Nunca use a posicao da linha dentro do arquivo .diff inteiro — o GitHub rejeita."
- Procure o hunk header
Colete todas as respostas. Cada agente retorna um array (vazio se nada a flagged).
7. Consolidar review draft
Crie $REVIEWS_DIR/PR-${PR_NUM}.md com este formato:
---
pr: <PR_NUM>
title: <titulo do PR>
author: <author>
branch: <head> -> <base>
stacks: <stacks detectadas>
status: draft
generated_at: <ISO timestamp>
---
# Review PR #<PR_NUM> — <titulo>
## Resumo
- **Arquivos analisados:** N
- **Stacks:** <lista>
- **Agentes executados:** <count>
- **Achados:** X critical, Y major, Z minor, W nit
## Comentarios inline
### `<path/do/arquivo>`
#### Linha <N> — <severity> [<stack>-<focus>]
<body do comentario>
#### Linha <M> — <severity> [<stack>-<focus>]
<body>
### `<outro/arquivo>`
...
## Veredicto sugerido
<APPROVE | COMMENT | REQUEST_CHANGES> — <justificativa em 1 linha>
Regra:
- Se ha qualquer `critical` ou >= 3 `major` → REQUEST_CHANGES
- Se ha `major` (1-2) ou apenas `minor`/`nit` → COMMENT
- Se nao ha achados → APPROVETambem salve o JSON consolidado para o /post-review:
# $REVIEWS_DIR/PR-${PR_NUM}.comments.json
[
{"path": "...", "line": N, "severity": "...", "body": "...", "agent": "stack-focus"},
...
]8. Apresentar para o usuario
Mostre um resumo conciso:
- Total de comentarios por severidade
- Veredicto sugerido
- Caminho do arquivo:
$REVIEWS_DIR/PR-${PR_NUM}.md
Finalize: "Review draft gerado. Revise/edite PR-<num>.md se quiser ajustar. Quando estiver pronto: /approve-review <PR_NUM>."
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

