feat: add grok, muse and posit; correct antigravity paths and verify zcode and pochi

MANIFEST-MATRIX.md

@@ -30,34 +30,34 @@ manifest and run the script — it updates every location automatically.
 
 | token | path | current_value |
 | --- | --- | --- |
-| agent_count | apps.json:18 | 87 |
-| agent_count | apps.json:19 | 87 |
-| agent_count | apps.json:40 | 87 |
-| agent_count | docs/agents.md:99 | 87 |
-| agent_count | docs/commands/agents.md:47 | 87 |
-| agent_count | docs/commands/agents.md:71 | 87 |
-| agent_count | docs/commands/agents.md:79 | 87 |
-| agent_count | docs/commands/doctor.md:25 | 87 |
-| agent_count | docs/commands/doctor.md:60 | 87 |
-| agent_count | docs/commands/doctor.md:77 | 87 |
-| agent_count | docs/comparison.md:10 | 87 |
-| agent_count | docs/guides/getting-started.md:18 | 87 |
-| verified_count | docs/guides/getting-started.md:18 | 27 |
-| agent_count | docs/index.md:7 | 87 |
-| verified_count | docs/index.md:7 | 27 |
-| agent_count | docs/index.md:38 | 87 |
-| verified_count | docs/index.md:38 | 27 |
-| agent_count | docs/migration-from-skills.md:10 | 87 |
-| agent_count | docs/migration-from-skills.md:54 | 87 |
-| agent_count | docs/reference.md:218 | 87 |
-| agent_count | package.json:4 | 87 |
-| agent_count | README.md:9 | 87 |
-| verified_count | README.md:9 | 27 |
-| verified_count | README.md:54 | 27 |
-| agent_count | README.md:90 | 87 |
-| agent_count | SKILL.md:5 | 87 |
-| verified_count | SKILL.md:5 | 27 |
-| agent_count | SKILL.md:10 | 87 |
-| verified_count | SKILL.md:10 | 27 |
-| agent_count | SKILL.md:86 | 87 |
-| verified_count | SKILL.md:86 | 27 |
+| agent_count | apps.json:18 | 90 |
+| agent_count | apps.json:19 | 90 |
+| agent_count | apps.json:40 | 90 |
+| agent_count | docs/agents.md:102 | 90 |
+| agent_count | docs/commands/agents.md:47 | 90 |
+| agent_count | docs/commands/agents.md:71 | 90 |
+| agent_count | docs/commands/agents.md:79 | 90 |
+| agent_count | docs/commands/doctor.md:25 | 90 |
+| agent_count | docs/commands/doctor.md:60 | 90 |
+| agent_count | docs/commands/doctor.md:77 | 90 |
+| agent_count | docs/comparison.md:10 | 90 |
+| agent_count | docs/guides/getting-started.md:18 | 90 |
+| verified_count | docs/guides/getting-started.md:18 | 34 |
+| agent_count | docs/index.md:7 | 90 |
+| verified_count | docs/index.md:7 | 34 |
+| agent_count | docs/index.md:38 | 90 |
+| verified_count | docs/index.md:38 | 34 |
+| agent_count | docs/migration-from-skills.md:10 | 90 |
+| agent_count | docs/migration-from-skills.md:54 | 90 |
+| agent_count | docs/reference.md:218 | 90 |
+| agent_count | package.json:4 | 90 |
+| agent_count | README.md:9 | 90 |
+| verified_count | README.md:9 | 34 |
+| verified_count | README.md:54 | 34 |
+| agent_count | README.md:90 | 90 |
+| agent_count | SKILL.md:5 | 90 |
+| verified_count | SKILL.md:5 | 34 |
+| agent_count | SKILL.md:10 | 90 |
+| verified_count | SKILL.md:10 | 34 |
+| agent_count | SKILL.md:86 | 90 |
+| verified_count | SKILL.md:86 | 34 |

README.md

@@ -6,7 +6,7 @@
 
 <p align="center">
   <b>The Security-First Skill Manager for AI Agents</b><br>
-   Static security scoring before any skill is installed · Skills & MCP Servers across 87 Agents (27 Verified)<br>
+   Static security scoring before any skill is installed · Skills & MCP Servers across 90 Agents (34 Verified)<br>
   Zero-dependency CLI · No signup · No marketplace required
 </p>
 
@@ -51,7 +51,7 @@
 
 
 <p align="center">
-  <b>⚡ Zero dependencies</b> · <b>🤖 27 verified agents</b> · <b>🔌 Skills + MCP</b> · <b>🔒 Security scoring</b> · <b>📝 Skill testing</b> · <b>🔧 Init templates</b> · <b>🌐 No marketplace required</b>
+  <b>⚡ Zero dependencies</b> · <b>🤖 34 verified agents</b> · <b>🔌 Skills + MCP</b> · <b>🔒 Security scoring</b> · <b>📝 Skill testing</b> · <b>🔧 Init templates</b> · <b>🌐 No marketplace required</b>
 </p>
 
 ---
@@ -87,7 +87,7 @@ rolecraft setup user/repo
 
 - **Zero dependencies** — only Node.js built-ins, nothing to audit
 - **Any source** — local folder, GitHub/GitLab/SSH URL, npm package
-- **87 agents** — opencode, claude-code, cursor, copilot, aider, oh-my-pi, and more
+- **90 agents** — opencode, claude-code, cursor, copilot, aider, oh-my-pi, and more
 - **No registry required** — works fully without a marketplace
 - **Security scoring** — static analysis before every install: prompt injection, command injection, obfuscated code, credential harvesting. Scores 0–100. It is a regex linter, not a policy engine — [what it does not catch](docs/security.md#what-this-does-not-catch)
 - **CI-ready** — lockfile-based re-install (`rolecraft ci`), `--yes` flag, `--dry-run`

SKILL.md

@@ -2,12 +2,12 @@
 name: rolecraft
 description: >-
   Install AI agent skills as roles & behaviors from any source — local folder,
-  GitHub, GitLab, SSH git URL. Zero-dependency CLI with 87 agent targets (27 verified).
+  GitHub, GitLab, SSH git URL. Zero-dependency CLI with 90 agent targets (34 verified).
 ---
 
 # rolecraft
 
-Install and manage AI agent skills across 87 agents (27 verified) with a single command. Zero dependencies. No registry required.
+Install and manage AI agent skills across 90 agents (34 verified) with a single command. Zero dependencies. No registry required.
 
 ## When to use
 
@@ -83,7 +83,7 @@ npx rolecraft convert ./my-skill
 | `--symlink` | Symlink instead of copy |
 | `--global` / `--project` | Scope selection |
 
-## Supported agents (87 total · 27 verified)
+## Supported agents (90 total · 34 verified)
 
 `opencode`, `claude-code`, `cursor`, `windsurf`, `devin`, `codex`, `copilot`, `aider`, `cline`, `gemini-cli`, `cody`, `continue`, `warp`, `codeium`, `fabric`, `goose`, `tabnine`, `supermaven`, `pr-pilot`, `loom`, `roo`, `trae`, `hermes`, `kiro`, `augment`, `kilo`, `openhands`, `junie`, `factory`, `command-code`, `cortex`, `mistral-vibe`, `qwen-code`, `openclaw`, `codebuddy`, `mux`, `pi`, `omp (oh-my-pi)`, `autohand-code`, `rovo-dev`, `firebender`, `ibm-bob`, `aider-desk`, and 45 more.
 

apps.json

@@ -15,8 +15,8 @@
       "id": "rolecraft",
       "name": "RoleCraft",
       "subtitle": "The package manager for AI agents",
-      "description": "Install, test, publish, and manage skills & MCP servers across 87 agents. Zero-dependency CLI.",
-      "longDescription": "RoleCraft is a zero-dependency Node.js CLI that serves as the package manager for AI agent skills. Install skills from any source, test them, publish to a central registry, and manage MCP servers — all from a single tool. Compatible with 87 AI coding agents including Claude Code, Cursor, Codex, OpenCode, Windsurf, Oh My Pi, and more.",
+      "description": "Install, test, publish, and manage skills & MCP servers across 90 agents. Zero-dependency CLI.",
+      "longDescription": "RoleCraft is a zero-dependency Node.js CLI that serves as the package manager for AI agent skills. Install skills from any source, test them, publish to a central registry, and manage MCP servers — all from a single tool. Compatible with 90 AI coding agents including Claude Code, Cursor, Codex, OpenCode, Windsurf, Oh My Pi, and more.",
       "icon": "https://raw.githubusercontent.com/rolecraft-sh/rolecraft/main/assets/rolecraft_logo.png",
       "iconEmoji": "📦",
       "iconStyle": {
@@ -37,7 +37,7 @@
         "Install skills from GitHub, URLs, or local paths",
         "Central skill registry with schema validation",
         "MCP server management — install, verify, and run",
-        "Compatible with 87 AI agents (Claude Code, Cursor, Codex, OpenCode, Oh My Pi...)",
+        "Compatible with 90 AI agents (Claude Code, Cursor, Codex, OpenCode, Oh My Pi...)",
         "GitHub Action for CI/CD integration",
         "Skill testing and verification built-in",
         "Profile system for team collaboration"

docs/agents.md

@@ -55,7 +55,7 @@ rolecraft knows where each AI agent looks for skills. When you use a flag like `
 | forge | `./.forge/skills/` | verified | - |
 | inference-sh | `~/.inferencesh/skills/` | experimental | - |
 | jazz | `~/.jazz/skills/` | verified | - |
-| iflow | `~/.iflow/skills/` | experimental | - |
+| iflow | `~/.iflow/skills/` | legacy | - |
 | kilo-code | `~/.kilocode/skills/` | experimental | - |
 | kode | `~/.kode/skills/` | experimental | - |
 | lingma | `~/.lingma/skills/` | verified | - |
@@ -70,21 +70,24 @@ rolecraft knows where each AI agent looks for skills. When you use a flag like `
 | zap | `~/.zap/skills/` | experimental | - |
 | codeep | `~/.codeep/skills/` | experimental | - |
 | kimi-code | `~/.kimi-code/skills/` | experimental | - |
-| zcode | `~/.zcode/skills/` | experimental | - |
+| zcode | `~/.zcode/skills/` | verified | - |
 | astrbot | `~/.astrbot/data/skills/` | experimental | - |
 | qoder-cn | `~/.qoder-cn/skills/` | community | - |
 | trae-cn | `~/.trae-cn/skills/` | experimental | - |
 | zenflow | `~/.zenflow/skills/` | experimental | - |
 | neovate | `~/.neovate/skills/` | experimental | - |
-| pochi | `~/.pochi/skills/` | experimental | - |
+| pochi | `~/.pochi/skills/` | verified | - |
 | adal | `~/.adal/skills/` | experimental | - |
 | droid | `~/.droid/skills/` | experimental | - |
 | chatgpt | `~/.agents/skills/` | verified | - |
 | codearts-agent | `~/.codeartsdoer/skills/` | experimental | - |
 | universal | `~/.config/agents/skills/` | experimental | - |
 | amp | `~/.config/agents/skills/` | verified | - |
-| antigravity | `~/.agents/skills/` | experimental | - |
-| antigravity-cli | `~/.agents/skills/` | experimental | - |
+| antigravity | `~/.gemini/config/skills/` | verified | - |
+| antigravity-cli | `~/.gemini/antigravity-cli/skills/` | verified | - |
+| grok | `~/.grok/skills/` | verified | - |
+| muse-code | `~/.config/muse/skills/` | verified | mcpServers |
+| posit-assistant | `~/.posit/assistant/skills/` | verified | - |
 | deep-agents | `~/.agents/skills/` | experimental | - |
 | dexto | `~/.agents/skills/` | experimental | - |
 | loaf | `~/.agents/skills/` | experimental | - |
@@ -94,9 +97,9 @@ rolecraft knows where each AI agent looks for skills. When you use a flag like `
 
 > **Support levels:** `verified` — actively tested and maintained; `community` — community-contributed, maintained on best-effort; `legacy` — previous generation, no active development; `experimental` — known to exist, not formally tested.
 
-> **MCP support:** 8 agent(s) support MCP server configuration. Format: `mcpServers`, `servers`, `experimental.mcpServers`
+> **MCP support:** 9 agent(s) support MCP server configuration. Format: `mcpServers`, `servers`, `experimental.mcpServers`
 
-> **Agent count:** 87 total — 27 verified, 4 community, 4 legacy, 52 experimental.
+> **Agent count:** 90 total — 34 verified, 4 community, 5 legacy, 47 experimental.
 
 ## Notes
 
@@ -123,9 +126,17 @@ rolecraft knows where each AI agent looks for skills. When you use a flag like `
 - **oh-my-pi:** pi fork by can1357; native skills at ~/.omp/agent/skills (user) and .omp/skills (project); also reads ~/.claude/skills, ~/.agents/skills and .github/skills; MCP via standard mcpServers format
 - **forge:** ForgeCode by tailcallhq; also reads ~/forge/skills and ~/.agents/skills; project dir has highest precedence
 - **jazz:** jazz-ai (lvndry/jazz, not an AWS product); also reads project ./skills
+- **iflow:** iFlow CLI was shut down on 2026-04-17; the vendor directs users to migrate to Qoder
 - **lingma:** Also reads project .lingma/skills; renamed to Qoder CN on 2026-05-20
+- **zcode:** ZCode, the GLM coding client. Reads ~/.zcode/skills/<name>/SKILL.md and ~/.agents/mcp.json
+- **pochi:** Pochi (TabbyML). Also reads ~/.agents/skills; project .pochi/skills
 - **chatgpt:** ChatGPT/Codex share the .agents skills locations; symlinked skill folders honored
 - **amp:** amp skill add --global installs to ~/.config/agents/skills; also reads ~/.agents/skills
+- **antigravity:** Antigravity 2.0 and the IDE read ~/.gemini/config/skills; the IDE also accepts legacy ~/.gemini/antigravity/skills. All surfaces share <workspace>/.agents/skills
+- **antigravity-cli:** Antigravity CLI (agy) has its own global dir, separate from the 2.0/IDE surface; plugin skills at ~/.gemini/antigravity-cli/plugins/<name>/skills. Shares <workspace>/.agents/skills
+- **grok:** Project skills at ./.grok/skills. Reads ~/.agents/skills via AGENTS.md compatibility. MCP lives in ~/.grok/config.toml as [mcp_servers.<name>] (TOML), which rolecraft does not yet write
+- **muse-code:** Muse Code by Meta. Scans ~/.config/muse/skills, ~/.agents/skills, ~/.claude/skills, ~/.codex/skills and project .agents/.codex/.claude/skills. MCP at ~/.config/muse/settings.json under mcpServers
+- **posit-assistant:** Posit Assistant (Positron, RStudio). Also reads ~/.agents/skills; project .posit/assistant/skills. Legacy ~/.positai/
 - **replit:** Project-committed .agents/skills only; no user-global skills dir
 - **zed:** Flat layout only: skills must be direct children of ~/.agents/skills
 

docs/commands/agents.md

@@ -44,7 +44,7 @@ $ rolecraft agents
 
 🔍 Agent Capability Manifest
 
-Total agents: 87
+Total agents: 90
 
 ## VERIFIED
 
@@ -68,15 +68,15 @@ These agents have built-in MCP configuration support:
 ## Validation
 
 Manifest valid: ✅ Yes
-Agent count: 87
+Agent count: 90
 ```
 
 ```bash
 $ rolecraft agents --json
 
 {
   "version": 1,
-  "agentCount": 87,
+  "agentCount": 90,
   "agents": [
     {
       "flag": "claude",

docs/commands/doctor.md

@@ -22,7 +22,7 @@ Scans your system and reports on:
 - **Lockfile schema** — validates global lockfile format
 - **Global & project lockfiles** — skill count per scope
 - **Disk usage** — total size of installed skills
-- **Agent detection** — finds installed agents among 87 supported, per-agent skill count and directory permissions
+- **Agent detection** — finds installed agents among 90 supported, per-agent skill count and directory permissions
 - **Orphaned skill dirs** — directories not tracked in any lockfile
 - **Skill integrity** — verifies skill directories exist, content hashes match, and checks for broken symlinks
 - **MCP servers** — counts configured MCP servers across detected agents
@@ -57,7 +57,7 @@ $ rolecraft doctor
    ✅ Global lockfile                        3 skill(s) tracked
    ⚠️  Project lockfile                       no project skills
    ✅ Disk usage                             3 skill(s), 12.5 KB total
-   ✅ Agent detection                        14/87 supported agents detected
+   ✅ Agent detection                        14/90 supported agents detected
      ✅  └ opencode                          2 skill(s) [rwx]
      ✅  └ claude-code                       1 skill(s) [rwx]
    ✅ Orphaned skill dirs                    none
@@ -74,7 +74,7 @@ $ rolecraft doctor --json
   "status": "degraded",
   "checks": {
     "Node.js version": { "status": "pass", "detail": "v22.0.0" },
-    "Agent detection": { "status": "pass", "detail": "14/87 supported agents detected" },
+    "Agent detection": { "status": "pass", "detail": "14/90 supported agents detected" },
     "Skill integrity": { "status": "warn", "detail": "3 checked, 1 missing director(ies)" },
     "MCP servers": { "status": "warn", "detail": "none configured" }
   },

docs/comparison.md

@@ -7,7 +7,7 @@
 | | rolecraft | skills (Vercel) | @agentskill.sh/cli |
 |---|---|---|---|
 | **Runtime** | Zero-dep Node ESM | 1 dep (Node) | 2 deps (TypeScript) |
-| **Agent targets** | **87** | 72 | 15+ |
+| **Agent targets** | **90** | 72 | 15+ |
 | **Skill marketplace** | skills.sh (90K+) | agentskill.sh (100K+) | — |
 | **Publish your own** | ❌ | ❌ | ❌ |
 | **MCP management** | ✅ | ❌ | ❌ |

docs/guides/getting-started.md

@@ -15,7 +15,7 @@ A skill is a reusable capability for your AI coding agent. Think of it as a plug
 - **Team conventions** — one command sets up every agent on every developer's machine
 - **MCP servers** — skills can declare and install MCP servers alongside themselves
 
-rolecraft manages all of this across **87 AI agents** (27 verified) with a single CLI and **zero runtime dependencies**.
+rolecraft manages all of this across **90 AI agents** (34 verified) with a single CLI and **zero runtime dependencies**.
 
 ---
 

docs/index.md

@@ -4,7 +4,7 @@ layout: home
 hero:
   name: "RoleCraft"
   text: "The Security-First Skill Manager for AI Agents"
-  tagline: Security scoring before install · Skills & MCP Servers across 87 Agents (27 Verified)
+  tagline: Security scoring before install · Skills & MCP Servers across 90 Agents (34 Verified)
   image:
     src: /rolecraft-demo.gif
     alt: RoleCraft Demo
@@ -35,4 +35,4 @@ features:
   - title: Init Templates
     details: "rolecraft init --template scaffolds production-ready skills from pre-built templates. Start fast, ship faster."
   - title: Parallel Install
-    details: Install skills and MCP servers across all 87 agents (27 verified) simultaneously. Blazing fast, built for scale.
+    details: Install skills and MCP servers across all 90 agents (34 verified) simultaneously. Blazing fast, built for scale.

docs/migration-from-skills.md

@@ -7,7 +7,7 @@ If you're using [Vercel's `skills` CLI](https://github.com/vercel-labs/skills) a
 | Reason | rolecraft | Vercel skills |
 |---|---|---|
 | Dependencies | **0** (zero-dep) | 1 (`supports-color`) |
-| Agent targets | **87** | 72 |
+| Agent targets | **90** | 72 |
 | Telemetry | **None** | Anonymous telemetry |
 | Offline installs | **Fully supported** | Requires network |
 | Shell completions | bash, zsh, fish | Not available |
@@ -51,7 +51,7 @@ rm ~/.skills-lock.json
 ## What rolecraft does differently
 
 ### More agent targets
-rolecraft supports **87 agents** vs Vercel's 72. This includes newer agents like `augment`, `kilo`, `openhands`, `junie`, `factory`, `command-code`, `oh-my-pi (omp)`, and more.
+rolecraft supports **90 agents** vs Vercel's 72. This includes newer agents like `augment`, `kilo`, `openhands`, `junie`, `factory`, `command-code`, `oh-my-pi (omp)`, and more.
 
 ### Any source, not just GitHub
 Install from local folders, GitLab, Bitbucket, SSH URLs, or even npm packages — not just GitHub repos.

docs/reference.md

@@ -215,7 +215,7 @@ Pass any of these to `install`, `setup`, or `profile` to target specific agents:
 | `--tiny-cloud` | tiny-cloud | `~/.tinycloud/skills/` |
 | `--zencoder` | zencoder | `~/.zencoder/skills/` |
 | `--codebuddy` | codebuddy | `~/.codebuddy/skills/` |
-| *(see [full list](agents) for all 87 agents)* | | |
+| *(see [full list](agents) for all 90 agents)* | | |
 
 Combine multiple flags in one command:
 

package.json

@@ -1,7 +1,7 @@
 {
   "name": "rolecraft",
   "version": "2.5.0",
-  "description": "The package manager for AI agents — install, test, publish, and manage skills & MCP servers across 87 agents. Zero-dependency CLI for claude-code, cursor, opencode, omp, and more.",
+  "description": "The package manager for AI agents — install, test, publish, and manage skills & MCP servers across 90 agents. Zero-dependency CLI for claude-code, cursor, opencode, omp, and more.",
   "type": "module",
   "bin": {
     "rolecraft": "bin/rolecraft.js"