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.
git clone --depth 1 https://github.com/felvieira/claude-skills-fvWrote 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.
[](https://agentmods.dev/commands/felvieira/claude-skills-fv/multi-plan)<a href="https://agentmods.dev/commands/felvieira/claude-skills-fv/multi-plan"><img src="https://agentmods.dev/badge/commands/felvieira/claude-skills-fv/multi-plan/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.
<a href="https://agentmods.dev/commands/felvieira/claude-skills-fv/multi-plan"><img src="https://agentmods.dev/badge/commands/felvieira/claude-skills-fv/multi-plan.svg" alt="Reviewed on agentmods" width="80" height="20"></a>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.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5.1 | $0.00033 | $0.00834 |
| Opus 5 | $0.00016 | $0.00417 |
| Sonnet 5 | $0.00007 | $0.00167 |
| Haiku 4.5 | $0.00003 | $0.00083 |
Grade A, and why
multi-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 12d 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.
What it actually says
/multi-plan — Disagreement Gate Multi-Modelo
Objetivo: Rodar o mesmo /plan em 2+ modelos diferentes (Claude + Codex), comparar outputs e surface só as discordâncias ao usuário pra decisão. Convergências são auto-aprovadas (alta confiança).
Quando usar:
- feature de alto risco (migração, segurança, arquitetura core)
- decisão de design controversa onde "1 modelo concorda consigo mesmo" é fraco demais
- antes de
/buildquando o plano precisa de validação externa - quando user explicitamente pediu "outra opinião" / "stress test"
Quando NÃO usar:
- task trivial (rename, lint, doc) — 1 modelo basta
- já tem spec validada via
/grill-meou/checklist - urgência de hotfix — overhead não compensa
- custo é fator dominante — multi-plan é ~2x mais caro
Inputs:
<descrição da feature>(obrigatório)--models=claude,codex(default) — qual subset rodar--depth=quick(default) — plano rápido |--depth=deep— plano completo com riscos
Output:
- planos paralelos lado a lado
- diff estruturado: convergências (auto-aprovadas) + divergências (AskUserQuestion)
- plano final = união aprovada
Protocolo:
-
Fan-out (skill 40 parallel-dispatcher):
- Worker A:
Agent(subagent_type='Plan', prompt='<descrição>')(claude) - Worker B:
Agent(subagent_type='codex:codex-rescue', prompt='<descrição> — produza plano de implementação')(codex) - (opcionais: outros modelos via subagent_type apropriado)
- Worker A:
-
Coleta dos N planos retornados (cada um tem: steps, riscos, files-to-edit, tests-needed)
-
Diff estruturado:
## Convergências (auto-aprovadas) - Step X: ambos concordam em editar arquivo Y - Risco Z: ambos identificam ## Divergências (decisão necessária) 1. Claude propõe estratégia A; Codex propõe B — qual? 2. Codex flag risco que Claude não viu — investigar? 3. Estimativa de esforço diverge — qual usar? -
AskUserQuestion apenas nas divergências (1 por turno, max 4).
-
Plano final = convergências + divergências resolvidas → entrega como output normal de
/plan.
Dependências:
- skill 40 (parallel-dispatcher) — fan-out paralelo
- subagent
codex:codex-rescue— execução em Codex policies/trade-off-resolution.md— quando os modelos discordam e user pede "qual usa", a hierarquia decidepolicies/model-routing.md— definição de qual modelo cada worker usa
Adaptado de: ECC /multi-plan (MIT). Nosso ganho vs original: integração com skill 40 + reuso de codex:rescue (não exige infra própria de multi-model).
Custo: ~2x tokens de /plan único. Use seletivamente.
Cross-refs: /plan, skill 40 (parallel-dispatcher), codex:rescue, policies/model-routing.md.
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.
- 12d ago First seen · 68 lines · 33 tokens per session scan A 5c1c49f86ebe
multi-plan is a command published in the GitHub repository felvieira/claude-skills-fv (23 stars, last pushed 2d ago), licensed Apache-2.0. It adds 33 tokens to every session and 834 once invoked, about $0.0002 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 2026-08-30.
Other commands, from other repositories
checklist
Generate a custom checklist for the current feature based on user requirements.
clarify
Identify underspecified areas in the current feature spec by asking up to 5 highly targeted clarification questions and encoding answers back into the spec.
specify
Create or update the feature specification from a natural language feature description.
analyze
Perform a non-destructive cross-artifact consistency and quality analysis across spec.md, plan.md, and tasks.md after task generation.
converge
Assess the current codebase against the feature's spec, plan, and tasks, then append any remaining unbuilt work as new tasks to tasks.md so implement can complete it.
implement
Execute the implementation plan by processing and executing all tasks defined in tasks.md.