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.
npx agentmods add commands/vynazevedo/first-plan/executegit clone --depth 1 https://github.com/vynazevedo/first-planWhat 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 | $0.00036 | $0.01745 |
| Opus 5 | $0.00018 | $0.00873 |
| Sonnet 5 | $0.00007 | $0.00349 |
| Haiku 4.5 | $0.00004 | $0.00175 |
Grade A, and why
execute 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 2d 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.
How it starts
The opening of the file, as written. The whole thing — 211 lines — stays where its author put it; the contents beside it link to each section on GitHub.
/fp:execute
Fase 3 - executa plano aprovado.
Argumentos
$ARGUMENTS:
- vazio: executa o plano ativo
--dry-run: simula sem aplicar mudanças, mostra diff conceitual
Workflow
Passo 1 - Pre-flight
- Ler
.first-plan/07-state/STATE.md - Verificar
phase: awaiting_approvalEactive_plandefinido- Se phase != awaiting_approval: rejeitar com explicação
- Se sem active_plan: rejeitar
- Ler
.first-plan/07-state/plans/<active_plan>.md - Ler
.first-plan/08-meta/questions.md:- Se há perguntas marcadas como críticas para este plano e ainda
answered: false: PARAR e listar
- Se há perguntas marcadas como críticas para este plano e ainda
Passo 2 - Snapshot pre-execute (v0.2.0+)
Antes de qualquer modificação, criar snapshot do estado atual de .first-plan/:
SNAPSHOT_ID="$(date -u +%Y-%m-%dT%H-%M-%S)-pre-execute-<slug>"
mkdir -p .first-plan/cache/snapshots/${SNAPSHOT_ID}
cp -r .first-plan .first-plan/cache/snapshots/${SNAPSHOT_ID}/snapshot 2>/dev/null
Habilita /fp:rollback se algo der errado.
Passo 3 - Atualizar STATE
.first-plan/07-state/STATE.md:
phase: executing- registrar timestamp de início
- registrar snapshot_id criado no Passo 2
Passo 4 - Confirmar premissas
Para cada arquivo que o plano marca como "modificar" ou "reusar":
- Verificar que o path existe
- Verificar que assinatura/símbolo referenciado bate
Se alguma premissa não bate:
- PARAR
phase: paused- Reportar ao usuário:
Premissa do plano falhou: <descrição> Evidência: <output do check> Opções: A) Replanejar: /fp:plan <feature> (refaz o plano com camada atualizada) B) Adaptar: descreva a alteração e prossiga C) Abortar: /fp:status (volta para discovered) - Aguardar instrução humana. Não improvisar.
Passo 5 - Executar passos na ordem
Para cada passo do plano:
- Mostrar ao usuário: "Passo X/N: <descrição>"
- Se
--dry-run: mostrar diff conceitual, não aplicar - Senão: aplicar via Edit/Write/MultiEdit
- Após aplicar:
- Verificar que aplicou conforme descrito
- Se houve linter/formatter automático no projeto, rodar inline (sem corrigir além do plano)
- Atualizar
STATE.mdcom último passo executado
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.
- 2d ago First seen · 211 lines · 36 tokens per session scan A f268aefd3fa8
execute is a command published in the GitHub repository vynazevedo/first-plan (23 stars, last pushed 8d ago), licensed MIT. It adds 36 tokens to every session and 1,745 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
gate
Run the full local gate (fmt, lint, test, build, docs, and fuzz when a parser changed) and report only what failed.
release
Prepare the next release per RELEASING.md — gate, price table, surface audit, changelog — and stop at the tag command.
matcha:why
🍵 Intent Discovery — problem, goals, success criteria, What → Why → How before touching code.
pr-review
Review a PR across multiple dimensions (description, code, scan, tests).
setup
Generate HOTL adapter files for your team's code assistants (Codex, Cline, Cursor, Copilot).
subagent-execute
Execute a reviewed HOTL workflow in this session with delegated subagent steps and controller-owned verification.