wize-dev-kit 0.14.0 → 0.15.0

package/AGENTS.md

@@ -2,7 +2,7 @@
 
 This repo is wired with the [`wize-dev-kit`](https://www.npmjs.com/package/wize-dev-kit).
 Detailed artifacts live under `.wize/`. The agents below are activated through your AI IDE
-using slash commands (Claude Code, Codex, Cursor, Windsurf, Antigravity all read this file).
+using slash commands (Claude Code, Codex, Cursor, Windsurf, Antigravity, Hermes all read this file).
 
 ## Operating context
 

package/CHANGELOG.md

@@ -9,6 +9,21 @@ Format inspired by [Keep a Changelog](https://keepachangelog.com/en/1.1.0/).
 
 ## [Unreleased]
 
+## [0.15.0] — 2026-09-02
+
+**Hermes Agent is now the 10th supported harness.** New `hermes` IDE target renders the kit's agents/skills/workflows as project-local skills at `.hermes/skills/wize-{code}/SKILL.md` (Anthropic-compatible format, same as Claude Code/Codex/Kimi Code). Hermes discovers project skills at the git root, loads them only when the repo is trusted (`hermes skills trust`), and lets trusted project skills override same-named profile skills — vendored repo skills win inside their repo. Hermes also reads root `AGENTS.md` as project context, and its CLI (`hermes -z "<prompt>"`) is registered in the installer's brownfield baseline for headless runs. Highlighted in all 3 READMEs + `docs/harnesses/hermes.md` (+pt-BR); registered in the installer targets, `doctor`, `.gitignore` block, and 4 test suites.
+
+### Added
+
+- **`hermes` IDE target** (`adapters/hermes/`) — renders via the shared Anthropic emitter to `.hermes/skills/`; companion files (`steps/`, `templates/`, `data/`) copied alongside `SKILL.md`. Selectable in `npx wize-dev-kit install` / `sync`; `install --yes` now enables 10 harnesses.
+- **`docs/harnesses/hermes.md` + `hermes.pt-BR.md`** — output layout, trust-gate setup (`hermes skills trust` / `skills.trusted_project_dirs`), precedence rules, and headless usage.
+
+### Changed
+
+- **Installer** — `TARGETS` gains `hermes`; `adapterTargetPath()` maps it to `.hermes/skills` (also covers `uninstall`); `.gitignore` block ignores `.hermes/skills/wize-*`; brownfield `baseline.js` detects the `hermes` CLI and runs `hermes -z` headlessly.
+- **Docs** — README (en/pt-BR/es) harness tables and summaries now list **10** IDE targets, with Hermes highlighted as new; `AGENTS.md` (repo + generated template) lists Hermes among the harnesses that read it.
+- **Tests** — adapter emit/companion coverage, structure, `install --yes` full-defaults, and `doctor` path mapping all now include `hermes`.
+
 ## [0.14.0] — 2026-08-31
 
 **`wize-check` — progress checklist with evidence.** A read-only orchestrator skill (owner: Wizer) that answers "where are we, what's done, what's left" mid-session without advancing or re-planning the work. The **evidence rule** is the core: an item only reaches *Done* with a commit, `file:line`, a passed command, a recorded gate, or a merged PR — intent and plans are never evidence. Done condition comes first: if it is not explicit, the skill states the assumed criterion and asks for confirmation instead of inventing a destination. After reporting, it resumes the interrupted work. Modes: default, `now`, `save` (`.wize/implementation/checks/`), `{topic}`. Registered in the orchestrator catalog, the intent routing table (Meta), all 9 harness adapters, READMEs (3 languages) and CHANGELOG. Suite green (539 tests, validate 80 files).

package/README.es.md

@@ -69,7 +69,7 @@ Tras instalar, abre tu IDE y di:
 
 ## Harnesses soportados
 
-Los 9 IDEs objetivo se renderizan desde la misma fuente; el formato y la mecánica varían por harness. **OpenCode** recibe la integración más profunda — la separación persona/workflow del kit se mapea sobre las primitivas propias de OpenCode (`mode: primary|subagent`, `agent:`, `subtask:`) en lugar de aplanarse en un único tipo de archivo.
+Los 10 IDEs objetivo se renderizan desde la misma fuente; el formato y la mecánica varían por harness. **OpenCode** recibe la integración más profunda — la separación persona/workflow del kit se mapea sobre las primitivas propias de OpenCode (`mode: primary|subagent`, `agent:`, `subtask:`) en lugar de aplanarse en un único tipo de archivo.
 
 | Harness | Salida | Destacado |
 |---|---|---|
@@ -77,6 +77,7 @@ Los 9 IDEs objetivo se renderizan desde la misma fuente; el formato y la mecáni
 | **Claude Code** | `.claude/skills/*/SKILL.md` | Formato Skill de Anthropic; fan-out ad-hoc vía Task/Agent tool (`wize-code-review`). [Docs →](docs/harnesses/claude-code.md) |
 | **Codex** | `.agents/skills/*/SKILL.md` | Mismo formato Skill + `AGENTS.md` en la raíz. [Docs →](docs/harnesses/codex.md) |
 | **Kimi Code** | `.kimi/skills/*/SKILL.md` | Mismo formato Skill; autodetecta árboles de skills de Claude/Codex. [Docs →](docs/harnesses/kimi-code.md) |
+| **Hermes Agent** 🆕 | `.hermes/skills/*/SKILL.md` | Mismo formato Skill, project-local; trust gate (`hermes skills trust`); las skills del proyecto sobreescriben las de perfil. [Docs →](docs/harnesses/hermes.md) |
 | **Antigravity** | `.agent/skills/*/SKILL.md` | Mismo formato Skill + `AGENTS.md` en la raíz. [Docs →](docs/harnesses/antigravity.md) |
 | **Cursor** | `.cursor/rules/*.mdc` | Reglas on-demand (`alwaysApply: false`), emparejadas por descripción. [Docs →](docs/harnesses/cursor.md) |
 | **Windsurf** | `.windsurf/rules/*.md` | Markdown plano; el modo de activación se define dentro del IDE. [Docs →](docs/harnesses/windsurf.md) |
@@ -255,7 +256,7 @@ npx wize-dev-kit version         # imprime la versión instalada del kit
 
 ## Estado
 
-**v0.12.1 — beta.** El método nunca produce estimaciones de desarrollo — ni horas, ni puntos, ni tallas de camiseta; una story se dimensiona solo por si cabe en un único PR. El ciclo completo (análisis → plan → solución → implementación) está montado con 10 agentes y una biblioteca estructurada de skills. Los releases recientes suman **contratos de misión** (`/wize-help mission`), **`wize-grill`** (entrevista-hasta-entendimiento antes de cualquier paso de autoría), **`wize-pre-pr-check`** (gate local de lint/format/build/tests unitarios antes de abrir un PR — cero costo de runner de CI) y **loop verification** en `wize-dev-story` (un loop de implementación auto-verificable con guarda de max-cycles que escala a `wize-correct-course`). El `security-overlay` (Pentester de IA) entrega un pipeline de pentest completo, un informe ejecutivo (puntuación de riesgo + briefing + plan de acción por IA) y planificación de remediación post-scan — validado de principio a fin contra una aplicación Laravel/PHP real. Los adapters de IDE para Claude Code, Cursor, Windsurf, Codex, Continue, Kimi Code, OpenCode y Antigravity se regeneran automáticamente — [OpenCode](docs/harnesses/opencode.md) recibe wiring nativo de `mode`/`agent`/`subtask`, la integración más profunda de las 9.
+**v0.12.1 — beta.** El método nunca produce estimaciones de desarrollo — ni horas, ni puntos, ni tallas de camiseta; una story se dimensiona solo por si cabe en un único PR. El ciclo completo (análisis → plan → solución → implementación) está montado con 10 agentes y una biblioteca estructurada de skills. Los releases recientes suman **contratos de misión** (`/wize-help mission`), **`wize-grill`** (entrevista-hasta-entendimiento antes de cualquier paso de autoría), **`wize-pre-pr-check`** (gate local de lint/format/build/tests unitarios antes de abrir un PR — cero costo de runner de CI) y **loop verification** en `wize-dev-story` (un loop de implementación auto-verificable con guarda de max-cycles que escala a `wize-correct-course`). El `security-overlay` (Pentester de IA) entrega un pipeline de pentest completo, un informe ejecutivo (puntuación de riesgo + briefing + plan de acción por IA) y planificación de remediación post-scan — validado de principio a fin contra una aplicación Laravel/PHP real. Los adapters de IDE para Claude Code, Cursor, Windsurf, Codex, Continue, Kimi Code, Hermes, OpenCode y Antigravity se regeneran automáticamente — [OpenCode](docs/harnesses/opencode.md) recibe wiring nativo de `mode`/`agent`/`subtask`, la integración más profunda de las 10.
 
 ---
 

package/README.md

@@ -69,7 +69,7 @@ After install, open your IDE and say:
 
 ## Supported harnesses
 
-All 9 IDE targets render from the same source; format and mechanics differ per harness. **OpenCode** gets the deepest integration — the kit's persona/workflow split maps onto OpenCode's own primitives (`mode: primary|subagent`, `agent:`, `subtask:`) instead of being flattened into one file type.
+All 10 IDE targets render from the same source; format and mechanics differ per harness. **OpenCode** gets the deepest integration — the kit's persona/workflow split maps onto OpenCode's own primitives (`mode: primary|subagent`, `agent:`, `subtask:`) instead of being flattened into one file type.
 
 | Harness | Output | Notable |
 |---|---|---|
@@ -77,6 +77,7 @@ All 9 IDE targets render from the same source; format and mechanics differ per h
 | **Claude Code** | `.claude/skills/*/SKILL.md` | Anthropic Skill format; ad-hoc Task/Agent-tool fan-out (`wize-code-review`). [Docs →](docs/harnesses/claude-code.md) |
 | **Codex** | `.agents/skills/*/SKILL.md` | Same Skill format + root `AGENTS.md`. [Docs →](docs/harnesses/codex.md) |
 | **Kimi Code** | `.kimi/skills/*/SKILL.md` | Same Skill format; auto-detects Claude/Codex skill trees. [Docs →](docs/harnesses/kimi-code.md) |
+| **Hermes Agent** 🆕 | `.hermes/skills/*/SKILL.md` | Same Skill format, project-local; trust gate (`hermes skills trust`); project skills override profile ones. [Docs →](docs/harnesses/hermes.md) |
 | **Antigravity** | `.agent/skills/*/SKILL.md` | Same Skill format + root `AGENTS.md`. [Docs →](docs/harnesses/antigravity.md) |
 | **Cursor** | `.cursor/rules/*.mdc` | On-demand rules (`alwaysApply: false`), matched by description. [Docs →](docs/harnesses/cursor.md) |
 | **Windsurf** | `.windsurf/rules/*.md` | Plain markdown; activation mode set inside the IDE. [Docs →](docs/harnesses/windsurf.md) |
@@ -250,7 +251,7 @@ npx wize-dev-kit version         # print the installed kit version
 
 ## Status
 
-**v0.12.1 — beta.** The method never produces development estimates — no hours, points, or t-shirt sizes; a story is sized only by whether it fits a single PR. The full lifecycle (analysis → plan → solution → implementation) is wired with 10 agents and a structured skill library. Recent releases add **mission contracts** (`/wize-help mission`), **`wize-grill`** (interview-to-shared-understanding before any authoring step), **`wize-pre-pr-check`** (local lint/format/build/unit-test gate before opening a PR — zero CI runner cost), and **loop verification** in `wize-dev-story` (a self-checking implementation loop with a max-cycles guard that escalates to `wize-correct-course`). The `security-overlay` (AI Pentester) ships a complete pentest pipeline, a stakeholder report (risk score + briefing + AI action plan), and post-scan remediation planning — validated end-to-end against a real Laravel/PHP app. IDE adapters for Claude Code, Cursor, Windsurf, Codex, Continue, Kimi Code, OpenCode, and Antigravity are regenerated automatically — [OpenCode](docs/harnesses/opencode.md) gets native `mode`/`agent`/`subtask` wiring, the deepest integration of the 9.
+**v0.12.1 — beta.** The method never produces development estimates — no hours, points, or t-shirt sizes; a story is sized only by whether it fits a single PR. The full lifecycle (analysis → plan → solution → implementation) is wired with 10 agents and a structured skill library. Recent releases add **mission contracts** (`/wize-help mission`), **`wize-grill`** (interview-to-shared-understanding before any authoring step), **`wize-pre-pr-check`** (local lint/format/build/unit-test gate before opening a PR — zero CI runner cost), and **loop verification** in `wize-dev-story` (a self-checking implementation loop with a max-cycles guard that escalates to `wize-correct-course`). The `security-overlay` (AI Pentester) ships a complete pentest pipeline, a stakeholder report (risk score + briefing + AI action plan), and post-scan remediation planning — validated end-to-end against a real Laravel/PHP app. IDE adapters for Claude Code, Cursor, Windsurf, Codex, Continue, Kimi Code, Hermes, OpenCode, and Antigravity are regenerated automatically — [OpenCode](docs/harnesses/opencode.md) gets native `mode`/`agent`/`subtask` wiring, the deepest integration of the 10.
 
 ---
 

package/README.pt-BR.md

@@ -69,7 +69,7 @@ Após instalar, abra sua IDE e diga:
 
 ## Harnesses suportadas
 
-Os 9 alvos de IDE são gerados a partir da mesma fonte; formato e mecânica mudam por harness. O **OpenCode** tem a integração mais profunda — a divisão persona/workflow do kit mapeia pras primitivas nativas do próprio OpenCode (`mode: primary|subagent`, `agent:`, `subtask:`) em vez de ser achatada num único tipo de arquivo.
+Os 10 alvos de IDE são gerados a partir da mesma fonte; formato e mecânica mudam por harness. O **OpenCode** tem a integração mais profunda — a divisão persona/workflow do kit mapeia pras primitivas nativas do próprio OpenCode (`mode: primary|subagent`, `agent:`, `subtask:`) em vez de ser achatada num único tipo de arquivo.
 
 | Harness | Saída | Destaque |
 |---|---|---|
@@ -77,6 +77,7 @@ Os 9 alvos de IDE são gerados a partir da mesma fonte; formato e mecânica muda
 | **Claude Code** | `.claude/skills/*/SKILL.md` | Formato Skill da Anthropic; fan-out ad hoc via Task/Agent tool (`wize-code-review`). [Docs →](docs/harnesses/claude-code.pt-BR.md) |
 | **Codex** | `.agents/skills/*/SKILL.md` | Mesmo formato Skill + `AGENTS.md` na raiz. [Docs →](docs/harnesses/codex.pt-BR.md) |
 | **Kimi Code** | `.kimi/skills/*/SKILL.md` | Mesmo formato Skill; autodetecta as árvores do Claude/Codex. [Docs →](docs/harnesses/kimi-code.pt-BR.md) |
+| **Hermes Agent** 🆕 | `.hermes/skills/*/SKILL.md` | Mesmo formato Skill, project-local; trust gate (`hermes skills trust`); skills do projeto sobrescrevem as de perfil. [Docs →](docs/harnesses/hermes.pt-BR.md) |
 | **Antigravity** | `.agent/skills/*/SKILL.md` | Mesmo formato Skill + `AGENTS.md` na raiz. [Docs →](docs/harnesses/antigravity.pt-BR.md) |
 | **Cursor** | `.cursor/rules/*.mdc` | Rules sob demanda (`alwaysApply: false`), casadas por descrição. [Docs →](docs/harnesses/cursor.pt-BR.md) |
 | **Windsurf** | `.windsurf/rules/*.md` | Markdown puro; modo de ativação definido dentro da IDE. [Docs →](docs/harnesses/windsurf.pt-BR.md) |
@@ -253,7 +254,7 @@ npx wize-dev-kit version         # imprime a versão instalada do kit
 
 ## Status
 
-**v0.12.1 — beta.** O método nunca produz estimativas de desenvolvimento — nada de horas, pontos ou tamanhos de camiseta; uma story é dimensionada apenas por caber (ou não) em um único PR. O ciclo completo (análise → plano → solução → implementação) está montado com 10 agentes e uma biblioteca estruturada de skills. Releases recentes adicionam **contratos de missão** (`/wize-help mission`), **`wize-grill`** (entrevista-até-entendimento antes de qualquer passo de autoria), **`wize-pre-pr-check`** (gate local de lint/format/build/testes unitários antes de abrir um PR — zero custo de runner de CI) e **loop verification** no `wize-dev-story` (loop de implementação auto-verificável com guarda de max-cycles que escala para `wize-correct-course`). O `security-overlay` (Pentester de IA) entrega um pipeline de pentest completo, um relatório executivo (score de risco + briefing + plano de ação por IA) e planejamento de correção pós-scan — validado de ponta a ponta contra uma aplicação Laravel/PHP real. Os adapters de IDE para Claude Code, Cursor, Windsurf, Codex, Continue, Kimi Code, OpenCode e Antigravity são regenerados automaticamente — o [OpenCode](docs/harnesses/opencode.pt-BR.md) tem a integração mais profunda das 9, com `mode`/`agent`/`subtask` nativos.
+**v0.12.1 — beta.** O método nunca produz estimativas de desenvolvimento — nada de horas, pontos ou tamanhos de camiseta; uma story é dimensionada apenas por caber (ou não) em um único PR. O ciclo completo (análise → plano → solução → implementação) está montado com 10 agentes e uma biblioteca estruturada de skills. Releases recentes adicionam **contratos de missão** (`/wize-help mission`), **`wize-grill`** (entrevista-até-entendimento antes de qualquer passo de autoria), **`wize-pre-pr-check`** (gate local de lint/format/build/testes unitários antes de abrir um PR — zero custo de runner de CI) e **loop verification** no `wize-dev-story` (loop de implementação auto-verificável com guarda de max-cycles que escala para `wize-correct-course`). O `security-overlay` (Pentester de IA) entrega um pipeline de pentest completo, um relatório executivo (score de risco + briefing + plano de ação por IA) e planejamento de correção pós-scan — validado de ponta a ponta contra uma aplicação Laravel/PHP real. Os adapters de IDE para Claude Code, Cursor, Windsurf, Codex, Continue, Kimi Code, Hermes, OpenCode e Antigravity são regenerados automaticamente — o [OpenCode](docs/harnesses/opencode.pt-BR.md) tem a integração mais profunda das 10, com `mode`/`agent`/`subtask` nativos.
 
 ---
 

package/adapters/README.md

@@ -12,6 +12,7 @@ Each adapter renders the kit's agents/skills/workflows into the file layout that
 | `codex` | `.agents/skills/wize-*/` | `SKILL.md` per agent/workflow | OpenAI Codex. |
 | `continue` | `.continue/prompts/wize-*.prompt` | `.prompt` | Continue prompt slot. |
 | `kimi-code` | `.kimi/skills/wize-*/` | `SKILL.md` per agent/workflow | Moonshot Kimi Code. |
+| `hermes` | `.hermes/skills/wize-*/` | `SKILL.md` per agent/workflow | Hermes Agent (project-local skills, Anthropic-compatible). |
 | `opencode` | `.opencode/agents/wize-*.md` + `.opencode/commands/wize-*.md` | Markdown | OpenCode CLI. Native agents + commands. |
 | `antigravity` | `.agent/skills/wize-*/` | `SKILL.md` per agent/workflow | Antigravity CLI + IDE. |
 | `generic` | `.wize/agents/wize-*.md` | Markdown | Fallback for any agent that can read a folder of markdown. |

package/adapters/hermes/README.md

@@ -0,0 +1,23 @@
+# Hermes Agent — Wize Dev Kit Adapter
+
+Emits the kit's agents/skills/workflows as Hermes project-local skills under
+`.hermes/skills/wize-{code}/SKILL.md`, using the Anthropic `SKILL.md` format
+that Hermes consumes verbatim (same as Claude Code / Codex / Kimi Code).
+
+## How Hermes picks these up
+
+- Hermes discovers project-local skills at `<git-root>/.hermes/skills/`
+  (Hermes-native) and `<git-root>/.agents/skills/` (cross-tool convention).
+- **Trust gate:** skills are only loaded when the repo root is trusted. Run
+  `hermes skills trust` (or add the repo to `skills.trusted_project_dirs` in
+  `~/.hermes/config.yaml`) once per checkout.
+- Trusted project skills **override** same-named profile/bundled skills —
+  vendored repo skills win inside their repo.
+- Hermes also reads root-level `AGENTS.md` (and `CLAUDE.md`, `.cursorrules`)
+  as project context, so installing the generic target alongside is a bonus,
+  not a requirement.
+
+## Files
+
+- `render.js` — renders via the shared `renderAnthropicSkills` emitter.
+- `adapter.yaml` — descriptor (target path, file pattern).

package/adapters/hermes/adapter.yaml

@@ -0,0 +1,8 @@
+code: hermes
+description: "Hermes Agent skills (.hermes/skills/, Anthropic-compatible, project-local)."
+target_path: ".hermes/skills/"
+file_pattern: "wize-*/SKILL.md"
+file_format: markdown
+hooks:
+  pre_render: []
+  post_render: []

package/adapters/hermes/render.js

@@ -0,0 +1,17 @@
+// Hermes Agent adapter — emits .hermes/skills/wize-{code}/SKILL.md per asset.
+// Hermes consumes the Anthropic SKILL.md format verbatim (same as Claude Code,
+// Codex, Kimi Code). Project-local skills live at <repo-root>/.hermes/skills/
+// and are only loaded when the repo root is trusted (`hermes skills trust`),
+// mirroring the OpenCode/Codex `.agents/skills/` convention.
+
+'use strict';
+
+const path = require('node:path');
+const { renderAnthropicSkills } = require('../../tools/installer/render-shared.js');
+
+function render(kitRoot, projectRoot, opts = {}) {
+  const targetDir = path.join(projectRoot, '.hermes', 'skills');
+  return renderAnthropicSkills(kitRoot, targetDir, opts);
+}
+
+module.exports = { render };

package/docs/harnesses/hermes.md

@@ -0,0 +1,26 @@
+# Hermes Agent — Wize Development Kit
+
+🌐 **Languages:** **English** · [Português (pt-BR)](hermes.pt-BR.md)
+
+← [Back to README](../../README.md)
+
+Hermes Agent consumes the same public Anthropic Skill format as Claude Code, Codex, and Kimi Code, as project-local skills.
+
+## Output
+
+`.hermes/skills/wize-{code}/SKILL.md` — one directory per persona, workflow, or skill, with companion files (`steps/`, `templates/`, `data/`) copied alongside.
+
+## Notable
+
+- **Project-local skills.** Hermes discovers `<git-root>/.hermes/skills/` (Hermes-native) and `<git-root>/.agents/skills/` (cross-tool convention) for sessions started inside that project.
+- **Trust gate.** Skills are load-on-demand procedure documents, so Hermes only loads project skills when the repo root is trusted — run `hermes skills trust` (or add the root to `skills.trusted_project_dirs` in `~/.hermes/config.yaml`) once per checkout. Untrusted repos are still discoverable with a one-line notice.
+- **Precedence.** Trusted project skills override same-named profile/bundled skills — vendored repo skills win inside their repo.
+- Hermes also reads root-level `AGENTS.md` (plus `CLAUDE.md`, `.cursorrules`) as project context, so pairing with the **Generic fallback** target is a bonus.
+
+## Setup
+
+Pick **Hermes Agent** as an IDE target during `npx wize-dev-kit install` (or add it and re-run `npx wize-dev-kit sync`). Run `hermes skills trust` in the repo, restart Hermes, then run `/wize-orchestrator`.
+
+## Headless
+
+`hermes -z "<prompt>"` runs a one-shot headless session, which the installer's brownfield baseline uses when Hermes is the detected harness CLI.

package/docs/harnesses/hermes.pt-BR.md

@@ -0,0 +1,26 @@
+# Hermes Agent — Wize Development Kit
+
+🌐 **Idiomas:** [English](hermes.md) · **Português (pt-BR)**
+
+← [Voltar ao README](../../README.pt-BR.md)
+
+O Hermes Agent consome o mesmo formato público de Skill da Anthropic usado pelo Claude Code, Codex e Kimi Code, na forma de skills project-local.
+
+## Saída
+
+`.hermes/skills/wize-{code}/SKILL.md` — um diretório por persona, workflow ou skill, com arquivos complementares (`steps/`, `templates/`, `data/`) copiados junto.
+
+## Destaques
+
+- **Skills project-local.** O Hermes descobre `<git-root>/.hermes/skills/` (nativo do Hermes) e `<git-root>/.agents/skills/` (convenção compartilhada) para sessões iniciadas dentro do projeto.
+- **Trust gate.** Skills são documentos de procedimento carregados sob demanda, então o Hermes só carrega skills do projeto quando a raiz do repositório é confiável — rode `hermes skills trust` (ou adicione a raiz em `skills.trusted_project_dirs` no `~/.hermes/config.yaml`) uma vez por checkout. Repositórios não confiáveis seguem detectáveis com um aviso de uma linha.
+- **Precedência.** Skills project-local confiáveis sobrescrevem skills de mesmo nome do perfil/bundle — skills versionadas no repo vencem dentro do próprio repo.
+- O Hermes também lê `AGENTS.md` na raiz (além de `CLAUDE.md`, `.cursorrules`) como contexto de projeto, então combinar com o alvo **Fallback genérico** é um bônus.
+
+## Como ativar
+
+Escolha **Hermes Agent** como alvo de IDE em `npx wize-dev-kit install` (ou adicione depois e rode `npx wize-dev-kit sync`). Rode `hermes skills trust` no repo, reinicie o Hermes e rode `/wize-orchestrator`.
+
+## Headless
+
+`hermes -z "<prompt>"` roda uma sessão headless de uma execução, usada pelo baseline brownfield do instalador quando o Hermes é o CLI de harness detectado.

package/package.json

@@ -1,7 +1,7 @@
 {
   "$schema": "https://json.schemastore.org/package.json",
   "name": "wize-dev-kit",
-  "version": "0.14.0",
+  "version": "0.15.0",
   "description": "Full-lifecycle AI-assisted development kit with Test Architect and Whiteport Design Studio embedded. Inspired by BMAD Method and WDS.",
   "keywords": [
     "ai",

package/tools/installer/baseline.js

@@ -5,7 +5,7 @@
 //
 // Harness priority is:
 //   1. user's selected IDE targets (from .wize/config/project.toml), highest first
-//   2. claude > codex > opencode (universal fallback)
+//   2. claude > codex > opencode > hermes (universal fallback)
 //
 // Set WIZE_SKIP_BASELINE=1 in the environment to disable any execution (the
 // install still suggests the next step). Useful for CI and unattended setups.
@@ -35,6 +35,11 @@ const HARNESSES = [
     code: 'opencode',
     binary: 'opencode',
     buildCmd: (prompt) => ({ cmd: 'opencode', args: ['run', prompt] })
+  },
+  {
+    code: 'hermes',
+    binary: 'hermes',
+    buildCmd: (prompt) => ({ cmd: 'hermes', args: ['-z', prompt] })
   }
 ];
 
@@ -96,7 +101,7 @@ function manualInstructions(harness) {
   if (!harness) {
     return [
       '',
-      'No AI harness CLI was detected on PATH (claude / codex / opencode).',
+      'No AI harness CLI was detected on PATH (claude / codex / opencode / hermes).',
       'Open your IDE in this repo and run:',
       '  /wize-document-project',
       ''

package/tools/installer/commands/doctor.js

@@ -60,6 +60,7 @@ function adapterTargetPath(targetCode, projectRoot) {
     case 'antigravity':  return path.join(projectRoot, '.agent/skills');
     case 'codex':        return path.join(projectRoot, '.agents/skills');
     case 'kimi-code':    return path.join(projectRoot, '.kimi/skills');
+    case 'hermes':       return path.join(projectRoot, '.hermes/skills');
     case 'cursor':       return path.join(projectRoot, '.cursor/rules');
     case 'windsurf':     return path.join(projectRoot, '.windsurf/rules');
     case 'continue':     return path.join(projectRoot, '.continue/prompts');

package/tools/installer/render-shared.js

@@ -286,7 +286,7 @@ function renderAgentsMd(kitRoot, projectRoot, opts = {}) {
     '',
     'This repo is wired with the [`wize-dev-kit`](https://www.npmjs.com/package/wize-dev-kit).',
     'Detailed artifacts live under `.wize/`. The agents below are activated through your AI IDE',
-    'using slash commands (Claude Code, Codex, Cursor, Windsurf, Antigravity all read this file).',
+    'using slash commands (Claude Code, Codex, Cursor, Windsurf, Antigravity, Hermes all read this file).',
     '',
     '## Operating context',
     '',

package/tools/installer/setup-helpers.js

@@ -24,6 +24,7 @@ const GITIGNORE_BODY = [
   '.agent/skills/wize-*',
   '.agents/skills/wize-*',
   '.kimi/skills/wize-*',
+  '.hermes/skills/wize-*',
   '.cursor/rules/wize-*.mdc',
   '.windsurf/rules/wize-*.md',
   '.continue/prompts/wize-*.prompt',

package/tools/installer/wize-cli.js

@@ -3,9 +3,9 @@
  * wize-dev-kit — CLI entry point
  * Subcommands: install, update, uninstall, list, sync, agent, workflow, help
  *
- * v0.14.0 — Full-lifecycle CLI with install, update, uninstall, list, sync,
+ * v0.15.0 — Full-lifecycle CLI with install, update, uninstall, list, sync,
  * agent, workflow, validate, doctor, and document-project commands. Adapters
- * render for 9 IDE targets. Security overlay ships with 8 tools.
+ * render for 10 IDE targets. Security overlay ships with 8 tools.
  * `install --yes` now enables all harnesses + all overlays by default.
  */
 'use strict';
@@ -37,6 +37,7 @@ const TARGETS = [
   { code: 'codex', label: 'Codex', default: false },
   { code: 'continue', label: 'Continue', default: false },
   { code: 'kimi-code', label: 'Kimi Code', default: false },
+  { code: 'hermes', label: 'Hermes Agent (.hermes/skills/)', default: false },
   { code: 'opencode', label: 'OpenCode', default: false },
   { code: 'antigravity', label: 'Antigravity (CLI + IDE)', default: false },
   { code: 'generic', label: 'Generic fallback (.wize/agents/)', default: true }