agentwheel 0.6.0 → 0.8.0

package/README.md

@@ -33,10 +33,10 @@ No lock-in. No central gatekeeper. Your packages live in plain git repos, your c
 
 ---
 
-> **Status: early (v0.6).** The lifecycle core is real and tested — local/git/skillkit/vercel
+> **Status: early (v0.7).** The lifecycle core is real and tested — local/git/skillkit/vercel
 > sources, optional registry discovery, plan/sync/update/drift/uninstall, overlays, eject/remember,
 > profiles, runtime auto-detection, fleet targeting, asset-includes, selective installs,
-> update notifications, rich JSON merge, and pluggable adapters.
+> update notifications, full Claude/Codex adapters, rich JSON/TOML merge, and pluggable adapters.
 > Expect sharp edges.
 
 ## What it does
@@ -101,14 +101,24 @@ For a control-plane setup, define named agents in config. Global config lives at
 ```jsonc
 {
   "agents": {
-    "lab-openclaw": { "adapter": "openclaw", "root": "/Users/me/.openclaw-home" },
-    "docs-copilot": { "adapter": "copilot", "root": "/Users/me/projects/docs" }
+    "lab-openclaw": { "adapter": "openclaw", "root": "/Users/me/.openclaw-home", "transport": "local" },
+    "remote-codex": {
+      "adapter": "codex",
+      "root": "/home/agent/project",
+      "transport": "ssh",
+      "host": "agent-host.example",
+      "user": "agent",
+      "port": 22,
+      "identityFile": "~/.ssh/id_ed25519"
+    }
   },
   "profiles": {
-    "daily": [
-      { "agent": "lab-openclaw" },
-      { "agent": "docs-copilot" }
-    ]
+    "daily": {
+      "runtimes": [
+        { "agent": "lab-openclaw" },
+        { "agent": "remote-codex" }
+      ]
+    }
   }
 }
 ```
@@ -119,6 +129,16 @@ agentwheel sync --all
 agentwheel sync --profile daily
 ```
 
+SSH targets use the same manifest and drift model as local targets. `plan --dry-run` reads the
+remote install manifest and hashes remote files before deciding whether a file is up to date,
+drifted, or conflicting. SSH hosts need `ssh`, `tar`, and `node` available on `PATH`.
+
+To scaffold a control-plane example:
+
+```bash
+agentwheel init --fleet-example
+```
+
 Target resolution order is exact: `--target-root` wins, then `--agent`, then auto-detect from the
 current directory, then fallback to the current directory.
 
@@ -139,9 +159,9 @@ Flow: **author + your workspace → `sync` → runtime**.
 A package is a git repo (or folder) with a JSON manifest and a canonical layout:
 
 ```jsonc
-// agentwheel.json  (plain JSON or JSONC — both work)
+// openpack.json  (plain JSON or JSONC — both work; agentwheel.json remains a legacy alias)
 {
-  "schemaVersion": 1,
+  "schemaVersion": 2,
   "name": "your-org/agent-pack",
   "version": "0.1.0",
   "provides": [
@@ -204,7 +224,7 @@ optional, and `agentwheel add <url|path>` always works without it.
 The public package registry lives at
 [`NestDevLab/agentwheel-registry`](https://github.com/NestDevLab/agentwheel-registry).
 
-1. Create a public repo with `agentwheel.json` and a standard layout such as `instructions/`, `rules/`, `skills/`, `commands/`, `mcp/`, or `hooks/`.
+1. Create a public repo with `openpack.json` and a standard layout such as `instructions/`, `rules/`, `skills/`, `commands/`, `mcp/`, or `hooks/`.
 2. Open a pull request to `agentwheel-registry` that adds an entry to `index.json`.
 3. Users install by short name:
 
@@ -260,6 +280,20 @@ agentwheel sync ./my-pack --adapter-config ./myco-internal.jsonc
 Built-in adapters ship for common runtimes; declarative adapters need no code and stay private.
 Programmatic adapters, for private runtime logic beyond file placement, require explicit `--allow-adapter-code`.
 
+Built-in runtime targets:
+
+| Runtime | Main targets |
+|---|---|
+| **OpenClaw** | `.openclaw/AGENTS.md`, `.openclaw/skills`, `.openclaw/rules`, `.openclaw/commands`, MCP/hooks/settings, semantic plugin planning |
+| **Claude Code** | `.claude/CLAUDE.md`, `.claude/skills`, `.claude/commands`, `.claude/agents`, `.claude/rules`, `.claude/.mcp.json`, `.claude/settings.json` |
+| **Codex CLI** | `.codex/AGENTS.md`, `.codex/skills`, `.codex/commands`, `.codex/agents`, `.codex/rules`, `.codex/config.toml`, `.codex/hooks.json` |
+| **Hermes** | `.hermes/AGENTS.md`, `.hermes/skills`, `.hermes/rules`, `.hermes/commands`, MCP/hooks/settings |
+| **GitHub Copilot** | `.github/copilot-instructions.md`, `.github/instructions`, `.github/prompts` |
+
+Claude MCP and hooks/settings are merged as JSON. Codex MCP entries are merged into
+`[mcp_servers]` in `.codex/config.toml` without deleting unrelated user config; hooks use
+`.codex/hooks.json`.
+
 Copilot support is intentionally file-drop only: instructions, rules, and prompt/command files are
 placed in GitHub-native locations, while raw `SKILL.md` directories stay disabled until there is a
 clear conversion format.
@@ -272,6 +306,7 @@ clear conversion format.
 - [x] **v0.4** — runtime auto-detection; no `--target-root` needed for normal use; fleet config with named agents; global + project config merge; `--agent` and `--all`.
 - [x] **v0.5** — asset-includes compose shared files into skills at install time; executable bits preserved; hashes include composed assets.
 - [x] **v0.6** — selective installs with `--select`/`--skill`; required artifacts; cached npm update notifier.
+- [x] **v0.7** — full Claude/Codex adapters; Codex TOML MCP merge; subagent enumeration; `--skill` selector fix.
 
 ## Design docs