feat(setup): add OpenCode integration

README.md

@@ -163,6 +163,17 @@ Kimi Code deploys the mnemon skill, prompt files, and native lifecycle hooks to
 `~/.kimi-code/` or `$KIMI_CODE_HOME/`. The integration registers
 `SessionStart`, `UserPromptSubmit`, and `Stop` hooks in `config.toml`.
 
+### [OpenCode](https://opencode.ai/)
+
+```bash
+mnemon setup --target opencode --yes
+```
+
+OpenCode deploys the mnemon skill to `.opencode/skills/`, registers the
+generated guide through `opencode.json` instructions, and installs a native
+plugin in `.opencode/plugins/`. The plugin injects recall context before chat
+requests and adds Mnemon guidance to session compaction.
+
 ### [OpenClaw](https://github.com/openclaw/openclaw)
 
 ```bash
@@ -263,7 +274,7 @@ memory is useful.
 
 - **Zero user-side operation** — install once; supported runtimes can use hooks, minimal runtimes can use persistent rules
 - **LLM-supervised** — the host LLM decides what to remember, update, and forget; no embedded LLM, no API keys
-- **Multi-framework support** — Claude Code, Codex, Cursor, TRAE/TRAE Work, Qoder/QoderWork, CodeBuddy, WorkBuddy, Kimi Code, and Hermes Agent (hooks), OpenClaw (plugins), Pi (extensions), Nanobot (skills), and more
+- **Multi-framework support** — Claude Code, Codex, Cursor, TRAE/TRAE Work, Qoder/QoderWork, CodeBuddy, WorkBuddy, Kimi Code, OpenCode, and Hermes Agent (hooks/plugins), OpenClaw (plugins), Pi (extensions), Nanobot (skills), and more
 - **Markdown-installable harness** — `SKILL.md`, `INSTALL.md`, `GUIDELINE.md`, and four lifecycle reminders
 - **Four-graph architecture** — temporal, entity, causal, and semantic edges, not just vector similarity
 - **Intent-native protocol** — three primitives (`remember`, `link`, `recall`) map to the LLM's cognitive vocabulary, not database syntax; structured JSON output with signal transparency
@@ -315,7 +326,7 @@ All your local agentic AIs — across sessions and frameworks — sharing one po
 
 The foundation is in place: a single `~/.mnemon` database that any agent can
 read and write. Claude Code, Codex, Cursor, TRAE/TRAE Work, Qoder/QoderWork,
-CodeBuddy, WorkBuddy, Kimi Code, and Hermes Agent setup automate hook installation;
+CodeBuddy, WorkBuddy, Kimi Code, OpenCode, and Hermes Agent setup automate hook/plugin installation;
 OpenClaw can use plugin hooks; Pi integrates via native skills and TypeScript
 lifecycle extensions; Nanobot integrates via skill files; NanoClaw integrates
 via container skills and volume mounts. The same harness can be installed in any

cmd/setup.go

@@ -22,11 +22,11 @@ var setupCmd = &cobra.Command{
 	Short: "Deploy mnemon into LLM CLI environments",
 	Long: `Detect installed LLM CLIs and deploy mnemon integration.
 
-By default, installs to project-local config (.claude/, .codex/, .cursor/, .trae/, .qoder/, .codebuddy/, .workbuddy/, .openclaw/, .nanobot/, .pi/).
-Use --global to install to user-wide config (~/.claude/, ~/.codex/, ~/.cursor/, ~/.trae/, ~/.qoder/, ~/.codebuddy/, ~/.workbuddy/, ~/.openclaw/, ~/.nanobot/workspace/, ~/.pi/agent/).
+By default, installs to project-local config (.claude/, .codex/, .cursor/, .trae/, .qoder/, .codebuddy/, .workbuddy/, .opencode/, .openclaw/, .nanobot/, .pi/).
+Use --global to install to user-wide config (~/.claude/, ~/.codex/, ~/.cursor/, ~/.trae/, ~/.qoder/, ~/.codebuddy/, ~/.workbuddy/, ~/.config/opencode/, ~/.openclaw/, ~/.nanobot/workspace/, ~/.pi/agent/).
 Hermes Agent, QoderWork, and Kimi Code use native user config at ~/.hermes/, ~/.qoderwork/, and ~/.kimi-code/.
 
-Supported environments: Claude Code, Codex, Cursor, Trae, Qoder, QoderWork, CodeBuddy, WorkBuddy, Kimi Code, OpenClaw, Nanobot, Pi, Hermes Agent.
+Supported environments: Claude Code, Codex, Cursor, Trae, Qoder, QoderWork, CodeBuddy, WorkBuddy, Kimi Code, OpenCode, OpenClaw, Nanobot, Pi, Hermes Agent.
 
 Examples:
   mnemon setup                              # Interactive: project-local install
@@ -39,6 +39,7 @@ Examples:
   mnemon setup --target codebuddy           # Non-interactive: CodeBuddy skill and hooks
   mnemon setup --target workbuddy           # Non-interactive: WorkBuddy skill and hooks
   mnemon setup --target kimi                # Non-interactive: Kimi Code skill and hooks
+  mnemon setup --target opencode            # Non-interactive: OpenCode skill and plugin
   mnemon setup --target hermes              # Non-interactive: Hermes Agent only
   mnemon setup --eject                      # Interactive: remove integrations
   mnemon setup --eject --target claude-code # Non-interactive: remove Claude Code only
@@ -47,16 +48,16 @@ Examples:
 }
 
 func init() {
-	setupCmd.Flags().StringVar(&setupTarget, "target", "", "target environment (claude-code, codex, cursor, trae, qoder, qoderwork, codebuddy, workbuddy, kimi, openclaw, nanobot, pi, hermes)")
+	setupCmd.Flags().StringVar(&setupTarget, "target", "", "target environment (claude-code, codex, cursor, trae, qoder, qoderwork, codebuddy, workbuddy, kimi, opencode, openclaw, nanobot, pi, hermes)")
 	setupCmd.Flags().BoolVar(&setupEject, "eject", false, "remove mnemon integrations")
 	setupCmd.Flags().BoolVar(&setupYes, "yes", false, "auto-confirm all prompts (CI-friendly)")
 	setupCmd.Flags().BoolVar(&setupGlobal, "global", false, "install to user-wide config instead of project-local config")
 	rootCmd.AddCommand(setupCmd)
 }
 
 func runSetup(cmd *cobra.Command, args []string) error {
-	if setupTarget != "" && setupTarget != "claude-code" && setupTarget != "codex" && setupTarget != "cursor" && setupTarget != "trae" && setupTarget != "qoder" && setupTarget != "qoderwork" && setupTarget != "codebuddy" && setupTarget != "workbuddy" && setupTarget != "kimi" && setupTarget != "openclaw" && setupTarget != "nanobot" && setupTarget != "pi" && setupTarget != "hermes" {
-		return fmt.Errorf("invalid target %q (must be claude-code, codex, cursor, trae, qoder, qoderwork, codebuddy, workbuddy, kimi, openclaw, nanobot, pi, or hermes)", setupTarget)
+	if setupTarget != "" && setupTarget != "claude-code" && setupTarget != "codex" && setupTarget != "cursor" && setupTarget != "trae" && setupTarget != "qoder" && setupTarget != "qoderwork" && setupTarget != "codebuddy" && setupTarget != "workbuddy" && setupTarget != "kimi" && setupTarget != "opencode" && setupTarget != "openclaw" && setupTarget != "nanobot" && setupTarget != "pi" && setupTarget != "hermes" {
+		return fmt.Errorf("invalid target %q (must be claude-code, codex, cursor, trae, qoder, qoderwork, codebuddy, workbuddy, kimi, opencode, openclaw, nanobot, pi, or hermes)", setupTarget)
 	}
 
 	envs := setup.DetectEnvironments(setupGlobal)
@@ -92,7 +93,7 @@ func runInstallFlow(envs []setup.Environment) error {
 
 	if len(detected) == 0 {
 		fmt.Println("\nNo supported LLM CLI environments detected.")
-		fmt.Println("Install Claude Code, Codex, Cursor, Trae, Qoder, QoderWork, CodeBuddy, WorkBuddy, Kimi Code, OpenClaw, Nanobot, Pi, or Hermes Agent, then run 'mnemon setup' again.")
+		fmt.Println("Install Claude Code, Codex, Cursor, Trae, Qoder, QoderWork, CodeBuddy, WorkBuddy, Kimi Code, OpenCode, OpenClaw, Nanobot, Pi, or Hermes Agent, then run 'mnemon setup' again.")
 		return nil
 	}
 
@@ -150,6 +151,8 @@ func installEnv(env *setup.Environment) error {
 		err = installWorkBuddy(env)
 	case "kimi":
 		err = installKimi(env)
+	case "opencode":
+		err = installOpenCode(env)
 	case "openclaw":
 		err = installOpenClaw(env)
 	case "nanobot":
@@ -902,6 +905,74 @@ func installKimi(env *setup.Environment) error {
 	return nil
 }
 
+// ─── OpenCode ───────────────────────────────────────────────────────
+
+func installOpenCode(env *setup.Environment) error {
+	configDir := env.ConfigDir
+
+	if !setupGlobal && !setupYes && setup.IsInteractive() {
+		home := setup.HomeDir()
+		localDir := ".opencode"
+		globalDir := home + "/.config/opencode"
+		idx := setup.SelectOne("Install scope",
+			[]string{
+				fmt.Sprintf("Local — this project only (%s/ + opencode.json)", localDir),
+				fmt.Sprintf("Global — all projects (%s/)", globalDir),
+			}, 0)
+		if idx == 1 {
+			configDir = globalDir
+		} else {
+			configDir = localDir
+		}
+	}
+
+	fmt.Printf("\nSetting up OpenCode (%s)...\n", configDir)
+
+	fmt.Println("\n[1/3] Skill")
+	if path, err := setup.OpenCodeWriteSkill(configDir); err != nil {
+		setup.StatusError(0, 0, "Skill", err)
+		return err
+	} else {
+		setup.StatusOK(0, 0, "Skill", path)
+	}
+
+	fmt.Println("\n[2/3] Prompts")
+	var promptPath string
+	if path, err := setup.WritePromptFiles(); err != nil {
+		setup.StatusError(0, 0, "Prompts", err)
+		return err
+	} else {
+		setup.StatusOK(0, 0, "Prompts", path)
+		promptPath = path
+	}
+	if path, err := setup.OpenCodeRegisterInstructions(configDir, promptPath); err != nil {
+		setup.StatusError(0, 0, "Instructions", err)
+		return err
+	} else {
+		setup.StatusUpdated(0, 0, "Instructions", path)
+	}
+
+	fmt.Println("\n[3/3] Plugin")
+	if path, err := setup.OpenCodeWritePlugin(configDir); err != nil {
+		setup.StatusError(0, 0, "Plugin", err)
+		return err
+	} else {
+		setup.StatusOK(0, 0, "Plugin", path)
+	}
+
+	fmt.Println()
+	fmt.Println("Setup complete!")
+	fmt.Printf("  Skill        %s/skills/mnemon/SKILL.md\n", configDir)
+	fmt.Printf("  Plugin       %s/plugins/mnemon.js\n", configDir)
+	fmt.Printf("  Instructions %s\n", setup.OpenCodeConfigPath(configDir))
+	fmt.Printf("  Prompts      %s/ (guide.md, skill.md)\n", promptPath)
+	fmt.Println()
+	fmt.Println("Restart OpenCode to activate the mnemon skill, instructions, and plugin.")
+	fmt.Println("Run 'mnemon setup --eject --target opencode' to remove.")
+
+	return nil
+}
+
 // ─── OpenClaw ───────────────────────────────────────────────────────
 
 func installOpenClaw(env *setup.Environment) error {
@@ -1378,6 +1449,13 @@ func ejectEnv(env *setup.Environment) error {
 			return errs[0]
 		}
 
+	case "opencode":
+		errs := setup.OpenCodeEject(env.ConfigDir)
+		ejectMarkdown("AGENTS.md", "Remove memory guidance from ./AGENTS.md?")
+		if len(errs) > 0 {
+			return errs[0]
+		}
+
 	case "openclaw":
 		errs := setup.OpenClawEject(env.ConfigDir)
 		ejectMarkdown("AGENTS.md", "Remove memory guidance from ./AGENTS.md?")

docs/USAGE.md

@@ -37,6 +37,7 @@ mnemon setup --target qoderwork
 mnemon setup --target codebuddy
 mnemon setup --target workbuddy
 mnemon setup --target kimi
+mnemon setup --target opencode
 mnemon setup --target openclaw
 mnemon setup --target pi
 mnemon setup --target nanobot --global
@@ -52,8 +53,8 @@ mnemon setup --eject --target claude-code
 
 | Flag | Default | Description |
 |---|---|---|
-| `--global` | `false` | Install to user-wide config instead of project-local (recommended for Nanobot: installs to `~/.nanobot/workspace/`; Pi installs to `~/.pi/agent/`; Hermes installs to `~/.hermes/`; QoderWork installs to `~/.qoderwork/`; Kimi Code installs to `~/.kimi-code/` or `$KIMI_CODE_HOME/`) |
-| `--target <name>` | (auto-detect) | Target environment: `claude-code`, `codex`, `cursor`, `trae`, `qoder`, `qoderwork`, `codebuddy`, `workbuddy`, `kimi`, `openclaw`, `nanobot`, `pi`, or `hermes` |
+| `--global` | `false` | Install to user-wide config instead of project-local (recommended for Nanobot: installs to `~/.nanobot/workspace/`; Pi installs to `~/.pi/agent/`; Hermes installs to `~/.hermes/`; QoderWork installs to `~/.qoderwork/`; Kimi Code installs to `~/.kimi-code/` or `$KIMI_CODE_HOME/`; OpenCode installs to `~/.config/opencode/`) |
+| `--target <name>` | (auto-detect) | Target environment: `claude-code`, `codex`, `cursor`, `trae`, `qoder`, `qoderwork`, `codebuddy`, `workbuddy`, `kimi`, `opencode`, `openclaw`, `nanobot`, `pi`, or `hermes` |
 | `--eject` | `false` | Remove mnemon integrations |
 | `--yes` | `false` | Auto-confirm all prompts |
 

docs/zh/README.md

@@ -143,6 +143,17 @@ Kimi Code 会将 mnemon skill、prompt 文件和原生生命周期 hooks 部署
 `~/.kimi-code/` 或 `$KIMI_CODE_HOME/`。该集成会在 `config.toml` 中注册
 `SessionStart`、`UserPromptSubmit` 和 `Stop` hooks。
 
+### [OpenCode](https://opencode.ai/)
+
+```bash
+mnemon setup --target opencode --yes
+```
+
+OpenCode 会将 mnemon skill 部署到 `.opencode/skills/`，通过
+`opencode.json` 的 `instructions` 注册生成的 guide，并在
+`.opencode/plugins/` 安装原生 plugin。该 plugin 会在聊天请求前注入
+recall context，并在 session compaction 中加入 Mnemon guidance。
+
 ### [OpenClaw](https://github.com/openclaw/openclaw)
 
 ```bash
@@ -226,7 +237,7 @@ Agent 工作，并且只在有用时调用 Mnemon
 
 - **零用户操作** — 安装一次；支持 hook 的 runtime 可用 hook，minimal runtime 可用持久规则
 - **LLM 监督式** — 宿主 LLM 主动决定记什么、更新什么、遗忘什么；无内嵌 LLM，无 API 密钥
-- **多框架支持** — Claude Code、Codex、Cursor、TRAE/TRAE Work、Qoder/QoderWork、CodeBuddy、WorkBuddy、Kimi Code 和 Hermes Agent（hooks）、OpenClaw（plugins）、Pi（extensions）、Nanobot（skills）等
+- **多框架支持** — Claude Code、Codex、Cursor、TRAE/TRAE Work、Qoder/QoderWork、CodeBuddy、WorkBuddy、Kimi Code、OpenCode 和 Hermes Agent（hooks/plugins）、OpenClaw（plugins）、Pi（extensions）、Nanobot（skills）等
 - **Markdown 可安装 harness** — `SKILL.md`、`INSTALL.md`、`GUIDELINE.md` 和四个生命周期提醒
 - **四图架构** — 时序、实体、因果、语义四种边，不仅仅是向量相似度
 - **意图原生协议** — 三个原语（`remember`、`link`、`recall`）映射到 LLM 的认知词汇而非数据库语法；结构化 JSON 输出，带信号透明度
@@ -275,7 +286,7 @@ Agent 工作，并且只在有用时调用 Mnemon
   Gemini CLI ───┘
 ```
 
-基础已就绪：一个 `~/.mnemon` 数据库，任何 agent 都可以读写。Claude Code、Codex、Cursor、TRAE/TRAE Work、Qoder/QoderWork、CodeBuddy、WorkBuddy、Kimi Code 和 Hermes Agent setup 可自动安装 hook；OpenClaw 可以使用 plugin hooks；Pi 通过原生 skill 和 TypeScript lifecycle extension 集成；Nanobot 通过 skill 文件集成；NanoClaw 通过容器技能和卷挂载集成。同一个 harness 可以安装到任何支持 skill、rule、system prompt 或 event hook 的 LLM CLI。
+基础已就绪：一个 `~/.mnemon` 数据库，任何 agent 都可以读写。Claude Code、Codex、Cursor、TRAE/TRAE Work、Qoder/QoderWork、CodeBuddy、WorkBuddy、Kimi Code、OpenCode 和 Hermes Agent setup 可自动安装 hook/plugin；OpenClaw 可以使用 plugin hooks；Pi 通过原生 skill 和 TypeScript lifecycle extension 集成；Nanobot 通过 skill 文件集成；NanoClaw 通过容器技能和卷挂载集成。同一个 harness 可以安装到任何支持 skill、rule、system prompt 或 event hook 的 LLM CLI。
 
 更长远的方向是**记忆网关**：协议层与存储引擎解耦。当前 SQLite 后端是第一个适配器；协议面（`remember / link / recall`）可运行在 PostgreSQL、Neo4j 或任何图数据库之上。Agent 侧优化（何时召回、记什么）与存储侧优化（索引、图算法）独立演进。详见[未来方向](design/08-decisions.md#82-未来方向)。
 

docs/zh/USAGE.md

@@ -37,6 +37,7 @@ mnemon setup --target qoderwork
 mnemon setup --target codebuddy
 mnemon setup --target workbuddy
 mnemon setup --target kimi
+mnemon setup --target opencode
 mnemon setup --target openclaw
 mnemon setup --target pi
 mnemon setup --target nanobot --global
@@ -52,8 +53,8 @@ mnemon setup --eject --target claude-code
 
 | 标志 | 默认值 | 说明 |
 |---|---|---|
-| `--global` | `false` | 安装到用户级配置而非项目本地（Nanobot 推荐安装到 `~/.nanobot/workspace/`；Pi 安装到 `~/.pi/agent/`；Hermes 安装到 `~/.hermes/`；QoderWork 安装到 `~/.qoderwork/`；Kimi Code 安装到 `~/.kimi-code/` 或 `$KIMI_CODE_HOME/`） |
-| `--target <name>` | (自动检测) | 目标环境：`claude-code`、`codex`、`cursor`、`trae`、`qoder`、`qoderwork`、`codebuddy`、`workbuddy`、`kimi`、`openclaw`、`nanobot`、`pi` 或 `hermes` |
+| `--global` | `false` | 安装到用户级配置而非项目本地（Nanobot 推荐安装到 `~/.nanobot/workspace/`；Pi 安装到 `~/.pi/agent/`；Hermes 安装到 `~/.hermes/`；QoderWork 安装到 `~/.qoderwork/`；Kimi Code 安装到 `~/.kimi-code/` 或 `$KIMI_CODE_HOME/`；OpenCode 安装到 `~/.config/opencode/`） |
+| `--target <name>` | (自动检测) | 目标环境：`claude-code`、`codex`、`cursor`、`trae`、`qoder`、`qoderwork`、`codebuddy`、`workbuddy`、`kimi`、`opencode`、`openclaw`、`nanobot`、`pi` 或 `hermes` |
 | `--eject` | `false` | 移除 mnemon 集成 |
 | `--yes` | `false` | 自动确认所有提示 |
 

internal/setup/assets/assets.go

@@ -107,6 +107,12 @@ var KimiUserPromptHook []byte
 //go:embed kimi/stop.sh
 var KimiStopHook []byte
 
+//go:embed opencode/SKILL.md
+var OpenCodeSkill []byte
+
+//go:embed opencode/mnemon.js
+var OpenCodePlugin []byte
+
 //go:embed openclaw/SKILL.md
 var OpenClawSkill []byte
 
@@ -157,5 +163,5 @@ var HermesCompactHook []byte
 
 // All returns the embedded filesystem for inspection.
 //
-//go:embed claude codex cursor trae qoder qoderwork codebuddy workbuddy kimi openclaw nanoclaw nanobot pi hermes
+//go:embed claude codex cursor trae qoder qoderwork codebuddy workbuddy kimi opencode openclaw nanoclaw nanobot pi hermes
 var All embed.FS

internal/setup/assets/opencode/SKILL.md

@@ -0,0 +1,46 @@
+---
+name: mnemon
+description: Persistent memory CLI for OpenCode. Store facts, recall past knowledge, link related memories, manage lifecycle.
+---
+
+# mnemon
+
+## Workflow
+
+1. **Remember**: `mnemon remember "<fact>" --cat <cat> --imp <1-5> --entities "e1,e2" --source agent`
+   - Diff is built in: duplicates are skipped, conflicts are auto-replaced.
+   - Output includes `action` (added/updated/skipped), `semantic_candidates`, and `causal_candidates`.
+2. **Link** (evaluate candidates from step 1 using judgment):
+   - Review `causal_candidates`: link only when the memories are genuinely causally related.
+   - Review `semantic_candidates`: high `similarity` alone is not enough; skip unrelated keyword matches.
+   - Syntax: `mnemon link <id> <candidate> --type <causal|semantic> --weight <0-1> [--meta '<json>']`
+3. **Recall**: `mnemon recall "<query>" --limit 10`
+
+## Commands
+
+```bash
+mnemon remember "<fact>" --cat <cat> --imp <1-5> --entities "e1,e2" --source agent
+mnemon link <id1> <id2> --type <type> --weight <0-1> [--meta '<json>']
+mnemon recall "<query>" --limit 10
+mnemon search "<query>" --limit 10
+mnemon import --dry-run <file>
+mnemon import <file>
+mnemon forget <id>
+mnemon related <id> --edge causal
+mnemon gc --threshold 0.4
+mnemon gc --keep <id>
+mnemon status
+mnemon log
+mnemon store list
+mnemon store create <name>
+mnemon store set <name>
+mnemon store remove <name>
+```
+
+## Guardrails
+
+- Use memory only when it can materially improve continuity or task quality.
+- Do not store secrets, passwords, tokens, private keys, or short-lived operational noise.
+- Categories: `preference`, `decision`, `insight`, `fact`, `context`
+- Edge types: `temporal`, `semantic`, `causal`, `entity`
+- Max 8,000 chars per insight.