@jaimevalasek/aioson 1.21.4 → 1.21.6

package/README.md

@@ -717,10 +717,11 @@ aioson agents
 aioson agent:prompt <agent> [--tool=codex|claude|gemini|opencode]
 aioson workflow:plan [path] [--classification=MICRO|SMALL|MEDIUM] [--json]
 aioson workflow:next [path] [--complete] [--auto-heal] [--force]
-aioson workflow:heal [path] --stage=<agent>
-aioson workflow:harden [path] [--dry-run]
-aioson workflow:execute [path] [--dry-run] [--start-from=<agent>]
-```
+aioson workflow:heal [path] --stage=<agent>
+aioson workflow:harden [path] [--dry-run]
+aioson workflow:execute [path] [--dry-run] [--start-from=<agent>]
+aioson intake:ask [path] --agent=<agent> --schema=<questions.json> [--out=<answers.json>] [--json]
+```
 
 </details>
 
@@ -985,12 +986,23 @@ aioson scan:project --json
 - [Runner system](docs/pt/5-referencia/runner-system.md)
 - [Secure by default](docs/pt/5-referencia/secure-by-default.md)
 
-*Living Memory*
-- [Memória viva](docs/pt/living-memory/memoria-viva.md)
-- [Autonomy contract](docs/pt/living-memory/autonomy-contract.md)
-
----
-
-## License
-
-[AGPL-3.0-only](LICENSE) — GNU Affero General Public License v3.0
+*Living Memory*
+- [Memória viva](docs/pt/living-memory/memoria-viva.md)
+- [Autonomy contract](docs/pt/living-memory/autonomy-contract.md)
+
+---
+
+## Research Inspirations
+
+AIOSON may study external agent-workflow patterns as product research. These links are inspiration references only: they are not runtime dependencies, endorsements, or citations inside generated AIOSON artifacts.
+
+- Interview-style clarification and lightweight decision records
+  - https://github.com/mattpocock/skills/tree/main/skills/productivity/grill-me
+  - https://github.com/mattpocock/skills/tree/main/skills/engineering/grill-with-docs
+  - https://github.com/mattpocock/skills/tree/main/README.md
+
+---
+
+## License
+
+[AGPL-3.0-only](LICENSE) — GNU Affero General Public License v3.0

package/bin/aioson.js

@@ -1,4 +1,4 @@
-#!/usr/bin/env node
+#!/usr/bin/env node
 'use strict';
 
 require('../src/cli.js');

package/docs/en/1-understand/ecosystem-map.md

@@ -71,8 +71,8 @@
 The default order depends on the classification:
 
 **MICRO:** `@setup → @product (optional) → @dev`
-**SMALL:** `@setup → @product → @sheldon (optional) → @analyst → @architect → @dev → @qa`
-**MEDIUM:** `@setup → @product → @sheldon → @analyst → @architect → @ux-ui → @pm → @orchestrator → @dev → @qa`
+**SMALL:** `@setup → @product → @sheldon (optional) → @analyst → @scope-check → @architect → @dev → @qa`
+**MEDIUM:** `@setup → @product → @sheldon → @analyst → @architect → @ux-ui → @pm → @orchestrator → @scope-check → @dev → @qa`
 
 > **Why does `@sheldon` appear so early?** It is the **PRD quality guardian** — it runs *between* `@product` and `@analyst` to detect gaps, validate technical assumptions with web research, and decide between enriching the PRD in-place or creating a phased plan in `.aioson/plans/{slug}/`. Can be invoked N times on the same PRD. Skipping this step on serious features is expensive later.
 

package/docs/en/1-understand/glossary.md

@@ -14,7 +14,7 @@ Terms in alphabetical order. Each entry has a **short definition** + **concrete
 
 **Example:** `@dev` is the agent that implements code. `@qa` is the one that writes tests. `@pentester` is the one that tries to break the feature on purpose.
 
-**How to invoke:** type `@name` in your AI client (Claude, Codex, Gemini, OpenCode) or describe the intent in natural language ("let's implement the feature").
+**How to invoke:** type `@name` in your AI client (Claude, Codex e OpenCode) or describe the intent in natural language ("let's implement the feature").
 
 ---
 
@@ -73,7 +73,7 @@ Terms in alphabetical order. Each entry has a **short definition** + **concrete
 
 **How it works:**
 - 0–1 points → **MICRO** (`@setup → @product → @dev`)
-- 2–3 points → **SMALL** (`@setup → @product → @analyst → @architect → @dev → @qa`)
+- 2–3 points → **SMALL** (`@setup → @product → @analyst → @scope-check → @architect → @dev → @qa`)
 - 4–6 points → **MEDIUM** (full workflow, all gates, all artifacts)
 
 **Where it appears:** `classification:` in the `project.context.md` frontmatter.
@@ -84,7 +84,7 @@ Terms in alphabetical order. Each entry has a **short definition** + **concrete
 
 **Definition:** the program where you actually talk to the AI. AIOSON is client-agnostic — it works with several.
 
-**Officially supported:** Claude Code, Codex CLI, Gemini CLI, OpenCode.
+**Officially supported:** Claude Code, Codex CLI, OpenCode.
 **Also works via terminal:** VS Code, Cursor, Windsurf, JetBrains IDEs, Zed, Antigravity (any that opens a terminal).
 
 **Where to learn more:** `5-reference/clientes-ai.md`.

package/docs/en/1-understand/what-is-aioson.md

@@ -56,7 +56,7 @@ your-project/
 │   ├── rules/               ← hard rules agents follow (security, etc.)
 │   ├── skills/              ← pluggable packages (design systems, processes)
 │   └── runtime/             ← local telemetry (SQLite)
-├── .claude/  .codex/  .gemini/  .windsurf/   ← integration with each AI client
+├── .claude/  .codex/                         ← native client config
 ├── CLAUDE.md  AGENTS.md  OPENCODE.md         ← per-client instructions
 └── docs/                                      ← optional documentation
 ```
@@ -67,7 +67,7 @@ Open your favorite AI client and type `@setup`, `@product`, `@dev`, etc. — the
 
 Works with **any IDE that has a terminal**:
 
-- Claude Code · Codex CLI · Gemini CLI · OpenCode
+- Claude Code · Codex CLI · OpenCode
 - VS Code, Google Antigravity, Cursor, Windsurf, JetBrains, Zed (with any of the clients above)
 
 Agents are *prompts*, not plugins. They live in `.md` files and your AI client reads them when you invoke via `@name`.

package/docs/en/1-understand/why-it-exists.md

@@ -31,7 +31,7 @@ You work on a feature for 2 hours with the AI, go to sleep, come back the next d
 
 ### Problem 3 — Inconsistency in the team
 
-In a team, each person has a different prompt style. Each AI client (Claude, Codex, Gemini) responds differently. The result is a codebase stitched together from different aesthetics and patterns.
+In a team, each person has a different prompt style. Each AI client (Claude, Codex) responds differently. The result is a codebase stitched together from different aesthetics and patterns.
 
 **How AIOSON solves it:**
 - The same agents (`.aioson/agents/`) are read by **any** AI client.

package/docs/en/2-start/existing-project.md

@@ -30,10 +30,9 @@ The interactive wizard appears:
 
 ```
 ? Which AI tools? ›
-   ◉ Claude Code
-   ◉ Codex CLI
-   ◯ Gemini CLI
-   ◯ OpenCode
+   ◉ Claude Code
+   ◉ Codex CLI
+   ◯ OpenCode
 
 ? Mode? ›
    ◉ Development
@@ -75,7 +74,7 @@ The scan will:
 ## Step 3 — Run `@setup`
 
 ```bash
-claude .   # or: codex . / gemini . / opencode .
+claude .   # or: codex . / opencode .
 ```
 
 Inside the AI client:

package/docs/en/2-start/first-project.md

@@ -11,7 +11,7 @@ We're going to build a **mini TODO app** using the SMALL cycle. You'll see each
 ## Prerequisites
 
 - **Node.js 18+** installed
-- An AI client configured: **Claude Code**, **Codex CLI**, **Gemini CLI**, or **OpenCode**. Examples here use Claude Code; switching is trivial.
+- An AI client configured: **Claude Code**, **Codex CLI**, or **OpenCode**. Examples here use Claude Code; switching is trivial.
 - A terminal open
 
 If you don't have an AI client yet: Claude Code is the most widely used by the AIOSON community. See [`clientes-ai.md`](../5-reference/clientes-ai.md).
@@ -52,7 +52,7 @@ my-todo/
 ## Step 2 — Open the AI client in the project
 
 ```bash
-claude .   # or: codex .   or: gemini .   or: opencode .
+claude .   # or: codex .   or: opencode .
 ```
 
 From here on, we'll interact with the AI.
@@ -296,7 +296,7 @@ It tells you who's next.
 | I want to resume an interrupted feature | Run `@deyvin` — it reads `dev-state.md` and continues. |
 | Not sure if SMALL is the right classification | Ask `@neo` — it explains the calculation. |
 | Install failed | `npx @jaimevalasek/aioson doctor` — diagnoses and suggests a fix. |
-| I want to add Codex/Gemini later | `npx @jaimevalasek/aioson install --reconfigure`. |
+| I want to add Codex later | `npx @jaimevalasek/aioson install --reconfigure`. |
 
 ---
 

package/docs/en/2-start/initial-decisions.md

@@ -41,7 +41,7 @@ The sum of three factors (each worth 0, 1, or 2 points):
 - Static portfolio page
 - Mini API with 3 endpoints
 
-#### SMALL — `@setup → @product → @analyst → @architect → @dev → @qa`
+#### SMALL — `@setup → @product → @analyst → @scope-check → @architect → @dev → @qa`
 
 - For: most real apps.
 - Full discovery+development+QA cycle.
@@ -88,7 +88,7 @@ You can mark **more than one** in the wizard — they coexist in the same projec
 |---|---|---|
 | **Claude Code** | Long agents, refactoring, planned tasks | Native skills, slash commands, hooks |
 | **Codex CLI** | Short tasks, focus on direct code | `@` mode to include files |
-| **Gemini CLI** | Multi-modal, low cost on some plans | Generous context window |
+| | Multi-modal, low cost on some plans | Generous context window |
 | **OpenCode** | Open-source, integrates with multiple providers | Granular configuration |
 
 **Recommendation for beginners:** start with Claude Code — it has the highest parity with AIOSON. Add others later with `aioson install --reconfigure`.
@@ -183,7 +183,7 @@ npx @jaimevalasek/aioson install --no-interactive
 
 | I want to change... | Command |
 |---|---|
-| Add Codex/Gemini to the same project | `aioson install --reconfigure` |
+| Add Codex to the same project | `aioson install --reconfigure` |
 | Activate Squads | `aioson install --reconfigure` (and select it) |
 | Switch design skill | `@ux-ui` in the AI client or `aioson install --reconfigure` |
 | Change interaction language | Edit `interaction_language:` in `project.context.md` or run `@setup` again |

package/docs/en/3-recipes/README.md

@@ -8,9 +8,9 @@ These three trails show **how features reach development** in AIOSON. You almost
 
 | Trail | When to use | Key agents |
 |---|---|---|
-| **[Full feature with @sheldon](./full-feature-with-sheldon.md)** | You have a PRD in mind and want the complete canonical trail (most used) | @product → @sheldon → @analyst → @architect → @ux-ui → @pm → @orchestrator → @dev → @qa |
+| **[Full feature with @sheldon](./full-feature-with-sheldon.md)** | You have a PRD in mind and want the complete canonical trail (most used) | @product → @sheldon → @analyst → @architect → @ux-ui → @pm → @orchestrator → @scope-check → @dev → @qa |
 | [From idea to PRD via @briefing](./from-idea-to-prd-via-briefing.md) | Your idea is still vague, several loose notes | @briefing → @product |
-| [External plans for @product](./external-plans-for-product.md) *(coming soon)* | You already planned in another chat (ChatGPT, Claude.io, Gemini Web) | @product (reads `/plans/`) |
+| [External plans for @product](./external-plans-for-product.md) *(coming soon)* | You already planned in another chat (ChatGPT, Claude.io Web) | @product (reads `/plans/`) |
 
 ## Recipes by scenario
 

package/docs/en/3-recipes/full-feature-with-sheldon.md

@@ -10,7 +10,7 @@
 ## The trail in one line
 
 ```
-/aioson:agent:product → @sheldon → @analyst → @architect → @ux-ui → @pm → @orchestrator → @dev
+/aioson:agent:product → @sheldon → @analyst → @architect → @ux-ui → @pm → @orchestrator → @scope-check → @dev
                                                               │
                                               (@qa ↔ @dev)  ← autonomous loop
                                                               │
@@ -299,7 +299,7 @@ Six months from now, anyone (or any AI) reads these files and understands **ever
 `@product → @dev → @qa`. Skip `@sheldon`, `@analyst`, `@architect`, `@ux-ui`, `@pm`, `@orchestrator`. The Constitution Article II (*Right-Sized Process*) protects you from unnecessary ceremony.
 
 ### SMALL without UI
-`@product → @sheldon → @analyst → @architect → @dev → @qa`. Skip `@ux-ui`, `@pm`, `@orchestrator`.
+`@product → @sheldon → @analyst → @scope-check → @architect → @dev → @qa`. Skip `@ux-ui`, `@pm`, `@orchestrator`.
 
 ### Full MEDIUM
 The complete trail above.

package/docs/en/5-reference/cli-reference.md

@@ -17,7 +17,7 @@ aioson init my-app --no-interactive
 
 **Options:**
 - `--lang=en|pt-BR|es|fr` — sets `conversation_language` in the generated context and applies the matching agent locale pack. Default: `en`.
-- `--tool=codex|claude|gemini|opencode` — configures the primary AI client. Affects which gateway file is used. Default: `codex`.
+- `--tool=codex|claude|opencode` — configures the primary AI client. Affects which gateway file is used. Default: `codex`.
 - `--no-interactive` — skip the wizard and install all files (CI / automation).
 - `--json` — prints structured JSON result instead of human-readable output.
 
@@ -44,8 +44,8 @@ aioson install --no-interactive
 
 **Options:**
 - `--lang=en|pt-BR|es|fr` — sets locale pack.
-- `--tool=codex|claude|gemini|opencode` — configures AI client.
-- `--reconfigure` — re-run the wizard even if a profile already exists (e.g. to add Gemini later).
+- `--tool=codex|claude|opencode` — configures AI client.
+- `--reconfigure` — re-run the wizard even if a profile already exists (e.g. to add another supported tool later).
 - `--no-interactive` — skip the wizard and install all files.
 - `--force` — overwrite existing files.
 - `--dry-run` — preview without writing.
@@ -129,7 +129,7 @@ aioson doctor --json
 
 **Checks performed:**
 - All files in `MANAGED_FILES` exist.
-- Gateway files (`CLAUDE.md`, `AGENTS.md`, `OPENCODE.md`, `.gemini/GEMINI.md`) are present.
+- Gateway files (`CLAUDE.md`, `AGENTS.md`, `OPENCODE.md`, `OPENCODE.md`) are present.
 - `.aioson/agents/` directory is populated.
 - `.aioson/context/` directory exists.
 
@@ -248,25 +248,25 @@ The locale shown reflects the active agent locale pack (from `project.context.md
 Print the activation prompt for a specific agent, ready to paste into any AI CLI that does not support slash commands.
 
 ```bash
-aioson agent:prompt setup
-aioson agent:prompt setup --tool=codex
-aioson agent:prompt ux-ui --tool=claude
-aioson agent:prompt dev --tool=gemini --json
+aioson agent:prompt setup
+aioson agent:prompt setup --tool=codex
+aioson agent:prompt ux-ui --tool=claude
+aioson agent:prompt dev --tool=opencode --json
 ```
 
 **Arguments:**
-- `<agent>` — agent id: `setup`, `product`, `analyst`, `architect`, `ux-ui`, `pm`, `dev`, `qa`, `orchestrator`.
+- `<agent>` — agent id: `setup`, `product`, `analyst`, `scope-check`, `architect`, `ux-ui`, `pm`, `dev`, `qa`, `orchestrator`.
 
 **Options:**
-- `--tool=codex|claude|gemini|opencode` — formats the prompt for the target CLI. Default: `codex`.
+- `--tool=codex|claude|opencode` — formats the prompt for the target CLI. Default: `codex`.
 - `--json` — returns structured JSON with the prompt string.
 
 **When to use:** if you're using an AI CLI that doesn't support `/aioson:agent:setup` slash commands, run this to get the exact text to paste into the chat.
 
 ```bash
-# Copy the prompt for @analyst in Gemini
-aioson agent:prompt analyst --tool=gemini
-# → paste the output into Gemini CLI
+# Copy the prompt for @analyst in OpenCode
+aioson agent:prompt analyst --tool=opencode
+# → paste the output into OpenCode
 ```
 
 ---
@@ -302,18 +302,18 @@ Notes:
 
 **Sequences by classification:**
 - `MICRO`: `@setup → @product (optional) → @dev`
-- `SMALL`: `@setup → @product → @analyst → @architect → @dev → @qa`
-- `MEDIUM`: `@setup → @product → @analyst → @architect → @ux-ui → @pm → @orchestrator → @dev → @qa`
+- `SMALL`: `@setup → @product → @analyst → @scope-check → @architect → @dev → @qa`
+- `MEDIUM`: `@setup → @product → @analyst → @architect → @ux-ui → @pm → @orchestrator → @scope-check → @dev → @qa`
 
 **Feature development workflow (after initial setup):**
 
 Once the project is set up, each new feature follows a shorter sequence — no `@setup` required:
 
 ```
-/aioson:agent:product → @analyst → @dev → @qa
+/aioson:agent:product → @analyst → @scope-check → @dev → @qa
 ```
 
-`@product` creates a feature-scoped `prd-{slug}.md` and registers the feature in `features.md`. `@analyst` produces `requirements-{slug}.md` and `spec-{slug}.md`. `@dev` reads the feature spec. `@qa` closes the feature by running `feature:close --verdict=PASS`, which updates `spec-{slug}.md` with a QA sign-off, marks it `done` in `features.md`, and automatically archives all feature artefacts to `.aioson/context/done/{slug}/`.
+`@product` creates a feature-scoped `prd-{slug}.md` and registers the feature in `features.md`. `@analyst` produces `requirements-{slug}.md` and `spec-{slug}.md`. `@scope-check` compares intent against the planned implementation before coding. `@dev` reads the feature spec. `@qa` closes the feature by running `feature:close --verdict=PASS`, which updates `spec-{slug}.md` with a QA sign-off, marks it `done` in `features.md`, and automatically archives all feature artefacts to `.aioson/context/done/{slug}/`.
 
 The `SMALL` and MEDIUM outputs include a note reminding you of this sequence.
 

package/docs/en/5-reference/mcp.md

@@ -33,7 +33,7 @@ aioson mcp:doctor --json
 Checks include:
 - `servers.local.json` existence and JSON validity
 - Core server baseline (`filesystem`, `context7`)
-- Preset coverage (`claude`, `codex`, `gemini`, `opencode`)
+- Preset coverage (`claude`, `codex`, `opencode`)
 - Required environment variables from enabled servers
 - Context compatibility for database and Web3 (`chain-rpc`)
 

package/docs/en/README.md

@@ -23,7 +23,7 @@ This is the entry door to the English documentation. It is not an alphabetical i
 ### I want a recipe ready for my case
 
 **Canonical trails — how features reach the dev:**
-1. **[Full feature with @sheldon](./3-recipes/full-feature-with-sheldon.md)** — the main trail: `@product → @sheldon → @analyst → @architect → @ux-ui → @pm → @orchestrator → @dev → @qa` (with optional gates from `@tester` and `@pentester`)
+1. **[Full feature with @sheldon](./3-recipes/full-feature-with-sheldon.md)** — the main trail: `@product → @sheldon → @analyst → @architect → @ux-ui → @pm → @orchestrator → @scope-check → @dev → @qa` (with optional gates from `@tester` and `@pentester`)
 2. [From idea to PRD via @briefing](./3-recipes/from-idea-to-prd-via-briefing.md) — when the idea is still vague
 3. [Continuity between sessions](./3-recipes/continuity-between-sessions.md) — feature dossier, dev-resume, drift detection
 
@@ -82,7 +82,7 @@ npx @jaimevalasek/aioson init my-project
 # 2. Enter the project
 cd my-project
 
-# 3. Open your AI client (Claude Code, Codex, Gemini, OpenCode) and type:
+# 3. Open your AI client (Claude Code, Codex e OpenCode) and type:
 /aioson:agent:setup
 ```
 

package/docs/en/deyvin-subtask-scout/README.md

@@ -21,7 +21,7 @@ This folder documents the full feature — 3 phases, 3 CLI verbs, 1 advisory che
 - Harness with sub-agent support: Claude Code (Agent tool) or Codex (MultiAgentV2)
 - `.aioson/config/scout-engine.json` in the project (copied automatically by the installer as `{}`, defaults active)
 
-> **Gemini CLI / OpenCode:** emit `harness_unsupported` and automatically fall back to `deyvin.md`'s CLI-less inline flow. Full harness parity is V2.
+> **OpenCode:** emit `harness_unsupported` and automatically fall back to `deyvin.md`'s CLI-less inline flow. Full harness parity is V2.
 
 ---
 

package/docs/en/deyvin-subtask-scout/sub-task-scout.md

@@ -85,7 +85,7 @@ Pure module `src/sub-task-engine.js`: prompt template builder, hand-rolled JSON
 Three CLI verbs (`scout:prep`, `scout:validate`, `scout:commit`) + state with file-lock in `src/sub-task-state.js`. Config template `template/.aioson/config/scout-engine.json` (empty `{}`; defaults active). Sandbox path check: `scope_paths` outside the project root are rejected.
 
 ### Phase 3 — `wiring-and-lifecycle`
-- `deyvin.md` (workspace + template byte-identical, 13611 bytes, under 15360 limit): new "Sub-task scout invocation" section with CLI path + CLI-less fallback per harness (Claude Code Agent tool, Codex MultiAgentV2, Gemini/OpenCode with `harness_unsupported` message)
+- `deyvin.md` (workspace + template byte-identical, 13611 bytes, under 15360 limit): new "Sub-task scout invocation" section with CLI path + CLI-less fallback per harness (Claude Code Agent tool, Codex MultiAgentV2/OpenCode with `harness_unsupported` message)
 - `feature:close`: archival hook copies scouts with matching `feature_slug` to `.aioson/context/features/{slug}/scouts/`, appends bullet to dossier
 - `memory:summary`: "Scouts dispatched: N (top topics: ...)" row always present — visible in agent cold-load bootstrap
 - `doctor`: `scouts_directory_pruning` advisory check (orphaned scouts >90d); `--fix` deletes them; scouts with `feature_slug` are **never** pruned

package/docs/en/deyvin-subtask-scout/troubleshooting.md

@@ -88,7 +88,7 @@ You have 1 retry (`max_retries_on_malformed_json=1`). If the second attempt also
 
 ## Harness without sub-agent support (`harness_unsupported`)
 
-**Symptom:** Gemini CLI or OpenCode don't natively support the Agent tool / sub-agent.
+**Symptom:** OpenCode don't natively support the Agent tool / sub-agent.
 
 **What happens:** `scout:prep` returns normally, but the harness emits `harness_unsupported` when trying to dispatch the sub-agent.
 

package/docs/pt/1-entender/glossario.md

@@ -14,7 +14,7 @@ Termos em ordem alfabética. Cada um tem **definição curta** + **exemplo concr
 
 **Exemplo:** `@dev` é o agente que implementa código. `@qa` é o que escreve testes. `@pentester` é o que tenta quebrar a feature de propósito.
 
-**Como invocar:** digite `@nome` no seu cliente AI (Claude, Codex, Gemini, OpenCode) ou descreva a intenção em linguagem natural ("vamos implementar a feature").
+**Como invocar:** digite `@nome` no seu cliente AI (Claude, Codex e OpenCode) ou descreva a intenção em linguagem natural ("vamos implementar a feature").
 
 ---
 
@@ -73,7 +73,7 @@ Termos em ordem alfabética. Cada um tem **definição curta** + **exemplo concr
 
 **Como funciona:**
 - 0–1 ponto → **MICRO** (`@setup → @product → @dev`)
-- 2–3 pontos → **SMALL** (`@setup → @product → @analyst → @architect → @dev → @qa`)
+- 2–3 pontos → **SMALL** (`@setup → @product → @analyst → @scope-check → @architect → @dev → @qa`)
 - 4–6 pontos → **MEDIUM** (workflow completo, todos os gates, todos os artefatos)
 
 **Onde aparece:** `classification:` no frontmatter do `project.context.md`.
@@ -84,7 +84,7 @@ Termos em ordem alfabética. Cada um tem **definição curta** + **exemplo concr
 
 **Definição:** o programa onde você de fato conversa com a IA. AIOSON é agnóstico — funciona com vários.
 
-**Suportados oficialmente:** Claude Code, Codex CLI, Gemini CLI, OpenCode.
+**Suportados oficialmente:** Claude Code, Codex CLI, OpenCode.
 **Funciona também via terminal:** VS Code, Cursor, Windsurf, JetBrains IDEs, Zed, Antigravity (qualquer um que abra um terminal).
 
 **Onde ver mais:** `clientes-ai.md`.

package/docs/pt/1-entender/mapa-do-ecossistema.md

@@ -71,8 +71,8 @@
 A ordem padrão depende da classificação:
 
 **MICRO:** `@setup → @product (opcional) → @dev`
-**SMALL:** `@setup → @product → @sheldon (opcional) → @analyst → @architect → @dev → @qa`
-**MEDIUM:** `@setup → @product → @sheldon → @analyst → @architect → @ux-ui → @pm → @orchestrator → @dev → @qa`
+**SMALL:** `@setup → @product → @sheldon (opcional) → @analyst → @scope-check → @architect → @dev → @qa`
+**MEDIUM:** `@setup → @product → @sheldon → @analyst → @architect → @ux-ui → @pm → @orchestrator → @scope-check → @dev → @qa`
 
 > **Por que `@sheldon` aparece tão cedo?** Ele é o **PRD quality guardian** — roda *entre* `@product` e `@analyst` para detectar gaps, validar premissas técnicas com pesquisa web, e decidir entre enriquecer o PRD in-place ou criar um phased plan em `.aioson/plans/{slug}/`. Pode ser invocado N vezes no mesmo PRD. Pular essa etapa em features sérias custa caro lá na frente.
 

package/docs/pt/1-entender/o-que-e-aioson.md

@@ -56,7 +56,7 @@ seu-projeto/
 │   ├── rules/               ← regras hard que agentes seguem (segurança, etc.)
 │   ├── skills/              ← pacotes plugáveis (design systems, processos)
 │   └── runtime/             ← telemetria local (SQLite)
-├── .claude/  .codex/  .gemini/  .windsurf/   ← integração com cada cliente AI
+├── .claude/  .codex/                         ← configuração nativa dos clientes
 ├── CLAUDE.md  AGENTS.md  OPENCODE.md         ← instruções por cliente
 └── docs/                                      ← documentação opcional
 ```
@@ -67,7 +67,7 @@ Você abre seu cliente AI favorito e digita `@setup`, `@product`, `@dev` etc.
 
 Funciona com **qualquer IDE que tenha um terminal**:
 
-- Claude Code · Codex CLI · Gemini CLI · OpenCode
+- Claude Code · Codex CLI · OpenCode
 - VS Code, Google Antigravity, Cursor, Windsurf, JetBrains, Zed (com qualquer um dos clientes acima)
 
 Os agentes são *prompts*, não plugins. Eles vivem em arquivos `.md` e o cliente AI os lê quando você invoca via `@nome`.

package/docs/pt/1-entender/por-que-existe.md

@@ -31,7 +31,7 @@ Você programa uma feature por 2 horas com a IA, vai dormir, volta no dia seguin
 
 ### Problema 3 — Inconsistência no time
 
-Em time, cada pessoa tem um estilo de prompt. Cada IA cliente (Claude, Codex, Gemini) responde diferente. O resultado é um codebase costurado por estéticas e padrões diferentes.
+Em time, cada pessoa tem um estilo de prompt. Cada IA cliente (Claude, Codex) responde diferente. O resultado é um codebase costurado por estéticas e padrões diferentes.
 
 **Como AIOSON resolve:**
 - Os mesmos agentes (`.aioson/agents/`) são lidos por **qualquer** cliente AI.

package/docs/pt/2-comecar/decisoes-iniciais.md

@@ -41,7 +41,7 @@ Soma de três fatores (cada um vale 0, 1 ou 2 pontos):
 - Página estática de portfólio
 - Mini-API de 3 endpoints
 
-#### SMALL — `@setup → @product → @analyst → @architect → @dev → @qa`
+#### SMALL — `@setup → @product → @analyst → @scope-check → @architect → @dev → @qa`
 
 - Para: a maioria dos apps reais.
 - Tem o ciclo completo de discovery+desenvolvimento+QA.
@@ -88,7 +88,7 @@ Você pode marcar **mais de um** no wizard — eles convivem no mesmo projeto.
 |---|---|---|
 | **Claude Code** | Agentes longos, refatorações, tarefas planejadas | Skills nativos, slash commands, hooks |
 | **Codex CLI** | Tarefas curtas, foco em código direto | Modo `@` para incluir arquivos |
-| **Gemini CLI** | Multi-modal, custo baixo em alguns planos | Janela de contexto generosa |
+| | Multi-modal, custo baixo em alguns planos | Janela de contexto generosa |
 | **OpenCode** | Open-source, integração com vários providers | Configuração granular |
 
 **Recomendação para iniciante:** comece com Claude Code, é o que tem mais paridade com AIOSON. Adicione outros depois com `aioson install --reconfigure`.
@@ -183,7 +183,7 @@ npx @jaimevalasek/aioson install --no-interactive
 
 | Quero mudar... | Comando |
 |---|---|
-| Adicionar Codex/Gemini ao mesmo projeto | `aioson install --reconfigure` |
+| Adicionar Codex ao mesmo projeto | `aioson install --reconfigure` |
 | Ativar Squads | `aioson install --reconfigure` (e marque) |
 | Trocar design skill | `@ux-ui` no cliente AI ou `aioson install --reconfigure` |
 | Mudar idioma de interação | Edite `interaction_language:` em `project.context.md` ou rode `@setup` de novo |

package/docs/pt/2-comecar/primeiro-projeto.md

@@ -11,7 +11,7 @@ Vamos construir um **mini-app de TODO** usando o ciclo SMALL. Você verá cada a
 ## Pré-requisitos
 
 - **Node.js 18+** instalado
-- Um cliente AI configurado: **Claude Code**, **Codex CLI**, **Gemini CLI** ou **OpenCode**. Os exemplos aqui usam Claude Code; é trivial trocar.
+- Um cliente AI configurado: **Claude Code**, **Codex CLI** ou **OpenCode**. Os exemplos aqui usam Claude Code; é trivial trocar.
 - Um terminal aberto
 
 Se você não tem cliente AI ainda: o Claude Code é o mais usado pela comunidade AIOSON. Veja [`clientes-ai.md`](../5-referencia/clientes-ai.md).
@@ -52,7 +52,7 @@ meu-todo/
 ## Passo 2 — Abrir o cliente AI no projeto
 
 ```bash
-claude .   # ou: codex .   ou: gemini .   ou: opencode .
+claude .   # ou: codex .   ou: opencode .
 ```
 
 A partir daqui, vamos conversar com a IA.
@@ -296,7 +296,7 @@ Ele te diz quem é o próximo.
 | Quero retomar uma feature interrompida | Rode `@deyvin` — ele lê `dev-state.md` e continua. |
 | Não sei se a classificação certa é SMALL | Pergunte ao `@neo` — ele explica o cálculo. |
 | Falhou ao instalar | `npx @jaimevalasek/aioson doctor` — diagnostica e sugere fix. |
-| Quero adicionar Codex/Gemini depois | `npx @jaimevalasek/aioson install --reconfigure`. |
+| Quero adicionar Codex depois | `npx @jaimevalasek/aioson install --reconfigure`. |
 
 ---
 

package/docs/pt/2-comecar/projeto-existente.md

@@ -30,10 +30,9 @@ O wizard interativo aparece:
 
 ```
 ? Quais ferramentas AI? ›
-   ◉ Claude Code
-   ◉ Codex CLI
-   ◯ Gemini CLI
-   ◯ OpenCode
+   ◉ Claude Code
+   ◉ Codex CLI
+   ◯ OpenCode
 
 ? Modo? ›
    ◉ Development
@@ -75,7 +74,7 @@ O scan vai:
 ## Passo 3 — Rodar `@setup`
 
 ```bash
-claude .   # ou: codex . / gemini . / opencode .
+claude .   # ou: codex . / opencode .
 ```
 
 Dentro do cliente AI:

package/docs/pt/3-receitas/README.md

@@ -8,9 +8,9 @@ Estas três trilhas mostram **como features chegam ao desenvolvimento** no AIOSO
 
 | Trilha | Quando usar | Agentes-chave |
 |---|---|---|
-| **[Feature completa com @sheldon](./feature-completa-com-sheldon.md)** | Você tem PRD em mente e quer a trilha canônica completa (a mais usada) | @product → @sheldon → @analyst → @architect → @ux-ui → @pm → @orchestrator → @dev → @qa |
+| **[Feature completa com @sheldon](./feature-completa-com-sheldon.md)** | Você tem PRD em mente e quer a trilha canônica completa (a mais usada) | @product → @sheldon → @analyst → @architect → @ux-ui → @pm → @orchestrator → @scope-check → @dev → @qa |
 | [Da ideia ao PRD via @briefing](./da-ideia-ao-prd-via-briefing.md) | Sua ideia ainda é vaga, várias anotações soltas | @briefing → @product |
-| [Plans externos para @product](./plans-externos-para-product.md) | Você já planejou em outro chat (ChatGPT, Claude.io, Gemini Web) | @product (lê `/plans/`) |
+| [Plans externos para @product](./plans-externos-para-product.md) | Você já planejou em outro chat (ChatGPT, Claude.io Web) | @product (lê `/plans/`) |
 
 ## Receitas por cenário
 

package/docs/pt/3-receitas/feature-completa-com-sheldon.md

@@ -10,7 +10,7 @@
 ## A trilha em uma linha
 
 ```
-@product → @sheldon → @analyst → @architect → @ux-ui → @pm → @orchestrator → @dev
+@product → @sheldon → @analyst → @architect → @ux-ui → @pm → @orchestrator → @scope-check → @dev
                                                               │
                                               (@qa ↔ @dev)  ← ciclo autônomo
                                                               │
@@ -299,7 +299,7 @@ Daqui a 6 meses, qualquer pessoa (ou IA) lê esses arquivos e entende **tudo**:
 `@product → @dev → @qa`. Pule `@sheldon`, `@analyst`, `@architect`, `@ux-ui`, `@pm`, `@orchestrator`. A Constitution Article II (*Right-Sized Process*) protege você de cerimônia desnecessária.
 
 ### SMALL sem UI
-`@product → @sheldon → @analyst → @architect → @dev → @qa`. Pule `@ux-ui`, `@pm`, `@orchestrator`.
+`@product → @sheldon → @analyst → @scope-check → @architect → @dev → @qa`. Pule `@ux-ui`, `@pm`, `@orchestrator`.
 
 ### MEDIUM puro
 A trilha completa acima.

package/docs/pt/3-receitas/landing-page.md

@@ -21,7 +21,7 @@ Este é um projeto **SMALL** (1 tipo de usuário, 0 integrações externas, sem
 ## Pré-requisitos
 
 - Node.js 18+ instalado
-- Claude Code (ou Codex/Gemini/OpenCode)
+- Claude Code (ou Codex/OpenCode)
 - AIOSON instalado no projeto (`npx @jaimevalasek/aioson init minha-landing`)
 - Opcional: uma URL de referência para estilo visual (qualquer site público que você admire)
 

package/docs/pt/5-referencia/README.md

@@ -39,7 +39,7 @@
 | Documento | Descrição |
 |---|---|
 | [comandos-cli.md](./comandos-cli.md) | Referência completa de todos os comandos do `aioson` |
-| [clientes-ai.md](./clientes-ai.md) | Como usar com Claude Code, Codex, Gemini CLI e OpenCode |
+| [clientes-ai.md](./clientes-ai.md) | Como usar com Claude Code, Codex e OpenCode |
 | [sandbox.md](./sandbox.md) | Execução segura com timeout, redação de secrets e output summarizado |
 | [compress-agents.md](./compress-agents.md) | Reduzir consumo de tokens comprimindo instruções de agentes |
 | [agent-sharding.md](./agent-sharding.md) | Carregar somente as seções relevantes de um agente para uma tarefa |