sdd-build

sdd-build is a command for Claude Code from leonardoaa/cloud-mcp. It costs 17 tokens per session (1,724 once invoked), scanned A, original, MIT.

A command for carrying out an approved software-development plan using SDD, a process that organizes work from a written development design. It also synchronizes subtasks with Jira and runs quality checks.

In plain words
What is it for?
Use it to execute planned development work, update related Jira subtasks, and perform QA.
Why use it?
It keeps the implementation, Jira task breakdown, and QA work aligned with an already approved plan.

Command for Claude Code

Written for Claude Code: allowed-tools in frontmatter.

Good fit Use it to execute planned development work, update related Jira subtasks, and perform QA.

Compare 6 commands from other repositories ↓
Install with agentmods
npx agentmods add commands/leonardoaa/cloud-mcp/sdd-build
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.

Clone the repo
git clone --depth 1 https://github.com/leonardoaa/cloud-mcp

Made for: Claude Code.

Wrote this? Show the measurements

A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.

agentmods badge for sdd-build

README.md
[![agentmods](https://agentmods.dev/badge/commands/leonardoaa/cloud-mcp/sdd-build/github.svg)](https://agentmods.dev/commands/leonardoaa/cloud-mcp/sdd-build)
Your own site
<a href="https://agentmods.dev/commands/leonardoaa/cloud-mcp/sdd-build"><img src="https://agentmods.dev/badge/commands/leonardoaa/cloud-mcp/sdd-build/github.svg" alt="Measured on agentmods" height="20"></a>

Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.

agentmods 80×15 button for sdd-build

Your own site · 80×15
<a href="https://agentmods.dev/commands/leonardoaa/cloud-mcp/sdd-build"><img src="https://agentmods.dev/badge/commands/leonardoaa/cloud-mcp/sdd-build.svg" alt="Reviewed on agentmods" width="80" height="20"></a>
Per session 17 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.
When invoked 1,724 The whole file, excluding the scripts and references it only reads on demand.
Security scan A 0 findings. A grade says what 26 rules found in the file — not that it is safe.
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 5.1 $0.00017 $0.01724
Opus 5.5 $0.00007 $0.00690
Sonnet 5.5 $0.00003 $0.00345
Haiku 4.5 $0.00002 $0.00172

Measured 2mo ago against content hash 8d6907bfdf28, method: parsed. Prices are Anthropic first-party input rates as of 2026-10-07, from the pricing page.

Security

Grade A, and why

sdd-build 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 2mo ago.

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.

resources/sdd/commands/sdd-build.md · 58 lines

How it starts

The opening of the file, as written. The whole thing — 58 lines — stays where its author put it; the contents beside it link to each section on GitHub.

Voce esta executando /sdd-build para implementar uma issue Jira planejada. Entrada: $ARGUMENTS.

Regras obrigatorias:

  1. Extraia exatamente uma chave PROJ-123. Antes do JIRA_GATE, execute a descoberta de peers do bloco multi-projeto: se a chave pertencer a um grupo, aplique o FANOUT e implemente cada membro, na order, executando as regras abaixo por completo dentro do workspace de cada membro; caso contrario, siga em MODO SINGLE sem alteracao. Em MODO SINGLE, execute o JIRA_GATE antes de ler ou alterar o workflow: resolva o vinculo; se ausente, permita escolha assistida, vincule e valide novamente. Sem Jira valido, encerre sem executar codigo.

  2. Leia a issue, confirme o projeto vinculado e somente entao abra constituicao e docs/sdd/specs/<ISSUE-KEY>/workflow.json.

  3. Recuse a execucao quando o workflow nao existir, nao registrar refinement.verdict: "PASS", nao possuir checklist aprovado, contiver NEEDS CLARIFICATION, estiver BLOCKED ou nunca tiver atingido READY_TO_BUILD; indique /sdd-plan <ISSUE-KEY>.

  4. Migre workflow v1 para v2 preservando dados. Antes de trabalhar, descarregue todos os pendingJiraEvents; falha impede retomada.

  5. Valide anexos e hashes. Mudanca material exige novo /sdd-plan.

  6. Capture buildStartedAt em ISO-8601 com timezone no momento de inicio real do build, persista no workflow.json e registre PHASE_STARTED no pai com eventKey <ISSUE>/<runId>/r<revision>/build/started, targetStatus: inProgress e o horario de inicio no resumo.

  7. Delegue exatamente sdd-orchestrator em modo BUILD. Para cada TASK-*, use exclusivamente subagent_type: "sdd-implementer"; para QA, sdd-qa-reviewer; nunca code ou fallback direto. Apos delegar a fase ao sdd-orchestrator, nao escreva workflow.json ate o retorno; a sequencia dos itens 8 a 13 e executada por quem estiver orquestrando.

  8. Para cada tarefa delegada, siga esta sequencia:

    Antes de delegar o agente:

    • Capture e persista startedAt na entrada da tarefa no workflow.json.
    • Registre TASK_STARTED na subtarefa com targetStatus: inProgress.

    Apos o retorno do agente:

    • Capture finishedAt e a observacao final.
    • Registre TASK_COMPLETED com arquivos, validacoes e targetStatus: done somente quando tudo passar.
  9. Tarefas consecutivas do mesmo lote marcadas [P] em tasks.md podem ser delegadas em paralelo, no maximo 3 simultaneas, cada uma em seu proprio worktree: registre TASK_STARTED e startedAt de todas antes de disparar o lote; integre os worktrees sequencialmente, na ordem das tarefas, somente depois de todo o lote retornar. Conflito de merge registra BUILD_BLOCKED preservando os worktrees. Se uma tarefa do lote falhar, aguarde o retorno das demais, registre os resultados individuais e so entao bloqueie. Em duvida sobre o isolamento, execute sequencial.

  10. Depois de cada retorno, consolide evidencias no workflow.json, incluindo horarios explicitos por tarefa. Se o agente criou worktree isolado, integre/mergeie as alteracoes aprovadas na branch de trabalho atual antes de marcar a tarefa como concluida. Nao repita tarefa concluida; bloqueios reais devem ser registrados na subtarefa correspondente.

  11. Em qualquer erro, capture buildFinishedAt em ISO-8601 com timezone, registre TASK_BLOCKED ou TASK_FAILED na subtarefa e BUILD_BLOCKED no pai; mantenha a subtarefa aberta e pare imediatamente. O comentario final do BUILD_BLOCKED deve informar Inicio do build: <buildStartedAt> e Fim do build: <buildFinishedAt>.

  12. Para QA, capture qa.startedAt, registre QA_STARTED na subtarefa e no pai; delegue exclusivamente sdd-qa-reviewer, informando o ciclo atual (qa.attempts + 1 de no maximo 3). Ao terminar, persista qa.finishedAt, qa.attempts, status e observacao.

  13. Em QA FAIL com correcoes minimas viaveis e dentro do escopo aprovado, execute um ciclo de correcao — no maximo 3 por build: incremente e persista qa.attempts; registre TASK_PROGRESS na subtarefa de QA com o ciclo N/3 e os achados; delegue sdd-implementer somente com as correcoes minimas apontadas pelo QA, com a mesma disciplina de worktree e merge do item 8; repita o QA completo. Esgotados os 3 ciclos com FAIL, ou quando a correcao exigir mudanca de escopo, registre QA_FAILED na subtarefa e BUILD_BLOCKED no pai e pare.

  14. QA_FAILED mantem cards abertos e bloqueia. Somente QA_PASSED conclui a subtask. 14b. Antes de BUILD_COMPLETED, confirme que todas as alteracoes dos worktrees dos agentes foram mergeadas na branch de trabalho atual e remova os worktrees temporarios criados para o build. Se algum worktree tiver alteracao nao consolidada, registre BUILD_BLOCKED e pare. Depois capture buildFinishedAt em ISO-8601 com timezone e persista no workflow.json.

  15. Registre BUILD_COMPLETED seguindo esta sequencia: a. Capture buildFinishedAt em ISO-8601 com timezone. Persista no workflow.json. b. Monte o resumo textual com:

    • Inicio do build: <buildStartedAt> e Fim do build: <buildFinishedAt>
    • Uma linha por TASK-* com status final e observacao curta
    • Uma linha de QA com status, ciclos executados e observacao
    • Validacoes executadas e observacoes gerais c. Chame jira_record_sdd_event no pai com eventType: BUILD_COMPLETED, targetStatus: done, o resumo montado em summary e as validacoes em validations. Nao inclua o campo report; o resumo em texto e suficiente. d. Em BUILD_BLOCKED, nao monte o resumo completo de build; registre apenas o evento com blockers.

Read the full file on GitHub · 58 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. 2mo ago First seen · 58 lines · 17 tokens per session scan A 8d6907bfdf28

Subscribe to this mod's changes

sdd-build is a command published in the GitHub repository leonardoaa/cloud-mcp (0 stars, last pushed 1mo ago), licensed MIT. It adds 17 tokens to every session and 1,724 once invoked, about $0.0001 per session on Opus 5.5. 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.