R
rmolinajr/MolBytes.Scaffold
Skill para Claude Code, Codex, Gemini CLI e Cursor que leva um projeto da ideia ao código testado, auditado e…
View on GitHub ↗Link checked 2026-11-12 · HTTP 200 · opens the vendor's own site
- Star count
- 1 ★ · rmolinajr/MolBytes.Scaffold · read from the GitHub search API · (checked 2026-11-13 20:00:50)
- Source
- read directly from its GitHub page · type: open-source project
- Harvest history
- first seen 2026-10-25 · last seen 2026-11-14 · seen 24× · listed by 2 sources
- Score
- score 203 · 1 ★ · 2 sources · repo
FAQ (answers come from our records)
Who is it for?
We filed it under these tags: Agent 技能. The tags come from the entry itself, not a taxonomy we imposed.
Who maintains it?
repository
rmolinajr/MolBytes.Scaffold;first catalogued 2026-10-25;last seen 2026-11-14。There is no official certification or quality rating here — we record sources and dates only.# MolBytes.Scaffold
以下摘录自该仓库的 README(片段) · 在 GitHub 读完整 README →
English · Português (Brasil)
> By MolBytes · created by Roberto Molina · project page
A skill for Claude Code, Codex, Gemini CLI
and Cursor that takes a software project from idea to tested, audited and
reviewed code, in 11 stages.
A main session acts as the orchestrator: it talks to you, hands each stage
to a specialized subagent running at the right model tier (light, medium or
strong; in Claude Code, haiku, sonnet or opus), checks the result and commits.
It never writes production code itself.
> Language: the skill's instructions, its in-pipeline commands
> (continuar, aprovado), its flags and the files it creates are in
> Brazilian Portuguese. You can describe your project in English.
## Pipeline
| # | Stage | Subagent | Output | Pauses for you |
| 1 | grill-me | main session | .project/qa.json | conversation |
| 2 | spec + threat model | scaffold-architect | .project/spec.md | approve spec |
| 3 | risks, tech debt, security requirements | scaffold-architect | .project/risks.yaml | — |
| 4 | design | scaffold-designer | .project/design/ | — |
| 5 | tickets | scaffold-planner | .project/tickets.yaml | — |
| 6 | model routing and waves | scaffold-router + scripts/ondas.py | .project/agent-map.json | approve plan and cost |
| 7 | parallel implementation in waves | scaffold-implementer | code | — |
| 8 | test plan → run → fix failures | scaffold-tester + scaffold-implementer | .project/plano-de-teste.yaml + tests | if a fix fails twice |
| 9 | security: audit → triage → fix → verify | scaffold-security + scaffold-implementer | .project/security/ | medium findings |
| 10 | code review | scaffold-reviewer | .project/code-review-report.md | decide on fixes |
| 11 | final tests + rescan + gate | scaffold-tester + scaffold-security | .project/pipeline-report.md | if blocked |
## Highlights
Resumable: state.md + git are the source of truth. At the pause
points, /clear and /molbytes:molbytes-scaffold continuar (the same
command works in a new session); after /compact, just continuar.
Extensible: an extension skill adds rules per stack, domain or
compliance without changing the pipeline; the contract is in
references/extensoes.md.
Changing an existing system: the skill looks for the system's
documentation (inside the project or elsewhere) and updates it along with
the code; without documentation, it only changes the system. Small changes
cost less: risks only for what changes, and design is skipped when there is
no new screen.
Code map: .project/mapa-codigo.md says where each thing lives
(structure, entry points, modules, contracts, commands), stamped with the
commit it describes. It is updated from the diff and checked at the gate,
so a new round starts from the map instead of rediscovering the code.
Round history: .project/historico.md keeps a short entry per round
(request, commits, areas, decisions, what was left out) and the changes
made outside the pipeline, so "when and why did this change?" has an
answer without opening each phase.
Documents outside the project, if you want: spec, design and reports
live in .project/ or in a separate folder with its own git, so the
client's project stays clean. See
INSTALL.md.
Parallelism without conflicts: tickets in the same wave never touch the
> By MolBytes · created by Roberto Molina · project page
A skill for Claude Code, Codex, Gemini CLI
and Cursor that takes a software project from idea to tested, audited and
reviewed code, in 11 stages.
A main session acts as the orchestrator: it talks to you, hands each stage
to a specialized subagent running at the right model tier (light, medium or
strong; in Claude Code, haiku, sonnet or opus), checks the result and commits.
It never writes production code itself.
> Language: the skill's instructions, its in-pipeline commands
> (continuar, aprovado), its flags and the files it creates are in
> Brazilian Portuguese. You can describe your project in English.
## Pipeline
| # | Stage | Subagent | Output | Pauses for you |
| 1 | grill-me | main session | .project/qa.json | conversation |
| 2 | spec + threat model | scaffold-architect | .project/spec.md | approve spec |
| 3 | risks, tech debt, security requirements | scaffold-architect | .project/risks.yaml | — |
| 4 | design | scaffold-designer | .project/design/ | — |
| 5 | tickets | scaffold-planner | .project/tickets.yaml | — |
| 6 | model routing and waves | scaffold-router + scripts/ondas.py | .project/agent-map.json | approve plan and cost |
| 7 | parallel implementation in waves | scaffold-implementer | code | — |
| 8 | test plan → run → fix failures | scaffold-tester + scaffold-implementer | .project/plano-de-teste.yaml + tests | if a fix fails twice |
| 9 | security: audit → triage → fix → verify | scaffold-security + scaffold-implementer | .project/security/ | medium findings |
| 10 | code review | scaffold-reviewer | .project/code-review-report.md | decide on fixes |
| 11 | final tests + rescan + gate | scaffold-tester + scaffold-security | .project/pipeline-report.md | if blocked |
## Highlights
Resumable: state.md + git are the source of truth. At the pause
points, /clear and /molbytes:molbytes-scaffold continuar (the same
command works in a new session); after /compact, just continuar.
Extensible: an extension skill adds rules per stack, domain or
compliance without changing the pipeline; the contract is in
references/extensoes.md.
Changing an existing system: the skill looks for the system's
documentation (inside the project or elsewhere) and updates it along with
the code; without documentation, it only changes the system. Small changes
cost less: risks only for what changes, and design is skipped when there is
no new screen.
Code map: .project/mapa-codigo.md says where each thing lives
(structure, entry points, modules, contracts, commands), stamped with the
commit it describes. It is updated from the diff and checked at the gate,
so a new round starts from the map instead of rediscovering the code.
Round history: .project/historico.md keeps a short entry per round
(request, commits, areas, decisions, what was left out) and the changes
made outside the pipeline, so "when and why did this change?" has an
answer without opening each phase.
Documents outside the project, if you want: spec, design and reports
live in .project/ or in a separate folder with its own git, so the
client's project stays clean. See
INSTALL.md.
Parallelism without conflicts: tickets in the same wave never touch the
Listed badge (optional)
Copy this line into the README if you want the listing mark. Optional, and it does not affect being listed.
[](https://anclum.com/en/s/rmolinajr-molbytes-scaffold/)
- Source
- github · https://github.com/rmolinajr/MolBytes.Scaffold
- Browse by field
- Agent skills
- Traction
- 1 ★
- Listed by
- 2 independent sources
- added
- today
More like this
Filed under
Every tag opens the full list for that filter. Counts are the live catalogue figures.