agentmods.dev

implement-plan

Instructions for carrying out an existing software-development plan through separate, sequential subagents. They define required workflow stages, reviews, file access, tests, and validation responsibilities.

In plain words
What is it for?
Use them when implementing an approved plan that requires worktrees, staged coding, code review, user-interface review, testing, or documentation.
Why use it?
They prevent an agent from skipping governance steps or making changes directly when the workflow requires controlled delegation. This makes execution easier to review and verify.

Skill for Claude CodeCodex

2 3mo ago A scan 19 tokens original MIT
View the source file Joaonic/agentkit
Install

Getting it into your agent

One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.

agentmods
npx agentmods add skills/joaonic/agentkit/implement-plan
Any agent
npx skills add Joaonic/agentkit --skill implement-plan
Clone the repo
git clone --depth 1 https://github.com/Joaonic/agentkit

Made for: Claude Code, Codex.

Per session 19 Skills are progressive disclosure: only the name and description are preloaded; the body loads when the skill is used.
When invoked 4,884 The whole file, excluding the scripts and references it only reads on demand.
Security scan A 0 findings. Scan, not verified.
Origin original No closer match found in the catalogue.
Token cost

What it costs to keep this loaded

Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.

ModelPer sessionOnce invoked
Fable 5 $0.00019 $0.04884
Opus 5 $0.00010 $0.02442
Sonnet 5 $0.00004 $0.00977
Haiku 4.5 $0.00002 $0.00488

Measured yesterday against content hash 58cb03486694, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

implement-plan scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured yesterday.

A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.

Nothing flagged

None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.

src/core/skills/implement-plan/SKILL.md · 253 lines

How it starts

The opening of the file, as written. The whole thing — 253 lines — stays where its author put it; the contents beside it link to each section on GitHub.

LEIA ESTA SKILL COMPLETAMENTE ANTES DE COMEÇAR

Implementar um plano existente. Não cria plano. Sempre numa worktree separada.

Fonte única:

  • docs/governance/cursor/workflow.md
  • docs/governance/cursor/workflow/*.md

Contrato operacional obrigatório desta skill:

  1. Você é o agente pai e atua apenas como orquestrador.
  2. Você não pode usar diretamente Read, Write, Edit, MultiEdit, Shell, Grep, Search, Browser, MCP, ferramentas de VCS remotas, Web, testes, ou qualquer outra tool fora da tool de subagent.
  3. Todo acesso a ficheiros, comandos, GitLab/GitHub (tracker), MCP, web, testes e validações deve acontecer dentro de subagents.
  4. A execução é obrigatoriamente sequencial. Uma etapa por vez.
  5. Cada chamada de subagent deve definir explicitamente subagent_type.
  6. Cada chamada de subagent deve incluir o marcador exato da etapa no prompt ou na description.
  7. Não pular etapas obrigatórias.
  8. Após workflow:code-review, cumprir sempre workflow:ux-review. Com alteração em fluxos ou superfície user-facing sob web/** (ou outro frontend definido no subprojeto), usar ux-reviewer. Sem alteração user-facing, usar verifier ou docs para registar «UX Review: não aplicável (sem frontend)» — não invocar ux-reviewer (rule 08-ux-mandatory.mdc).
  9. O ux-reviewer (e a etapa workflow:ux-review) não substituem workflow:posttask.
  10. Antes de declarar concluído, cumprir a skill posttask, MR pareado (GitLab) ou PR (exceção GitHub), review-open-pr e CI verde conforme o workflow do projeto.

Política de zero tolerância (obrigatória em todas as etapas):

  1. Não existe "aprovado com ressalvas" — qualquer problema encontrado em scope (código, testes, UX, docs, acessibilidade) é BLOQUEANTE e deve ser resolvido antes de avançar.
  2. Se o code-reviewer ou ux-reviewer reportar problemas, o agente pai deve delegar correção ao subagent adequado e re-submeter para review até zero problemas restarem.
  3. Problemas de UX em ficheiros tocados são bloqueantes quando há mudança user-facing em web/** (ou frontend do subprojeto em causa).

Uso obrigatório de MCP para pesquisa:

  1. Subagents de workflow:investigation, workflow:implementation e workflow:code-review devem usar Context7 (servidor MCP configurado no projeto) para validar APIs/bibliotecas contra documentação actualizada (Spring Boot, React, Next.js, Flyway, MapStruct, etc.).
  2. Quando existirem outros MCPs oficiais em .cursor/mcp.json para integrações em scope, usá-los em vez de suposições de treino.
  3. Nunca depender apenas de training data para APIs/bibliotecas quando MCP aplicável estiver disponível.

Regras obrigatórias de CI (aplicar em todas as etapas que envolvam push/CI):

  1. CI RED é BLOQUEANTE — nenhuma etapa avança com checks falhados. Corrigir antes de prosseguir.
  2. CI CANCELLED == CI FAILED — pipeline fail-fast cancela jobs satélite. Nunca reexecutar pipelines sem corrigir causa raiz. Em GitLab: glab ci list --ref <branch>, depois glab ci trace <job-id> no job que falhou primeiro. Em GitHub Actions (ex.: web/your-github-project): gh run view <id> --json jobs e gh run view <id> --log-failed.
  3. Bugs encontrados durante code-review ou posttask são BLOQUEANTES — o agente pai delega correção e re-submete até zero problemas.

Regras obrigatórias de rebase/merge (NUNCA perder código):

  1. Ao fazer rebase ou merge de main na branch de trabalho, NUNCA perder conteúdo que já existia em main. O rebase serve para incorporar main — não para sobrescrever.
  2. Durante resolução de conflitos, ambos os lados devem ser preservados: o código novo da feature E o código existente de main. Se houver conflito real (mesmo trecho modificado nos dois lados), o agente deve fundir manualmente, mantendo a semântica de ambos.
  3. Após rebase, o agente deve verificar: git diff origin/main..HEAD -- <ficheiros com conflito> e confirmar que nenhuma funcionalidade, import, teste, config ou bloco de código de main desapareceu.
  4. Se o agente detectar que perdeu código de main durante rebase, deve imediatamente corrigir (cherry-pick, re-apply, ou editar manualmente) antes de avançar para qualquer outra etapa.
  5. Esta regra aplica-se a todas as etapas: investigation, implementation, code-review, posttask — sempre que houver rebase/merge de main.

Read the full file on GitHub · 253 lines

Changes

What this file has done since we first saw it

Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.

  1. yesterday First seen · 253 lines · 19 tokens per session scan A 58cb03486694

Subscribe to this mod's changes

implement-plan is a skill published in the GitHub repository Joaonic/agentkit (2 stars, last pushed 3mo ago), licensed MIT. It adds 19 tokens to every session and 4,884 once invoked, about $0.0001 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.

Code reviewTestingWriting and docs
Related

Other skills, from other repositories

systematic-debugging

Use when encountering any bug, test failure, or unexpected behavior, before proposing fixes.

obra/superpowers · 21 tokens

next-cache-components-adoption

Turn on Cache Components in a Next.js app and resolve the blocking routes it surfaces. Use when the user wants to enable, adopt, or migrate to Cache Components, flip the cacheComponents flag, work through a flood of blocking-prerender / instant validation errors, run the cache-components-instant-false codemod, or…

vercel/next.js · 95 tokens

babysit-pr

Babysit a GitHub pull request after creation by continuously polling review comments, CI checks/workflow runs, and mergeability state until the PR is merged/closed or user help is required. Diagnose failures, retry likely flaky failures up to 3 times, auto-fix/push branch-related issues when appropriate, and keep…

openai/codex · 114 tokens

imagegen

Generate or edit raster images when the task benefits from AI-created bitmap visuals such as photos, illustrations, textures, sprites, mockups, or transparent-background cutouts. Use when Codex should create a brand-new image, transform an existing image, or derive visual variants from references, and the output…

openai/codex · 113 tokens

cpu-profile-analysis

Analyze V8/Chrome CPU profiles (.cpuprofile) and DevTools trace files (Trace-.json). Use when: profiling performance, investigating slow functions, comparing code paths, finding bottlenecks, analyzing timeToRequest, understanding call trees from sampling profiler data, analyzing layout/paint/rendering, investigating…

microsoft/vscode · 71 tokens

next-cache-components-optimizer

Drive a Next.js route to instant navigation by setting up an agentic loop, under Cache Components / PPR, on initial load (hard navigation) and client-side navigation (soft navigation). Encode the goal as a failing @next/playwright instant() e2e and work it to green, one verified route at a time; the shipped test then…

vercel/next.js · 170 tokens