mcp-agent-foundry 2.1.0 → 2.2.0
This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
- package/README.md +52 -2
- package/package.json +1 -1
package/README.md
CHANGED
|
@@ -34,13 +34,34 @@ Agent Foundry extends Claude Code with multi-provider AI orchestration and git w
|
|
|
34
34
|
|
|
35
35
|
## Quick Start
|
|
36
36
|
|
|
37
|
-
###
|
|
37
|
+
### Option 1: Claude Code Plugin (Recommended)
|
|
38
|
+
|
|
39
|
+
Install as a Claude Code plugin for slash commands and auto-invoked skills:
|
|
40
|
+
|
|
41
|
+
```bash
|
|
42
|
+
# Add the marketplace
|
|
43
|
+
claude plugin marketplace add github:sashabogi/agent-foundry
|
|
44
|
+
|
|
45
|
+
# Install the plugin
|
|
46
|
+
claude plugin install agent-foundry
|
|
47
|
+
```
|
|
48
|
+
|
|
49
|
+
Restart Claude Code, then use:
|
|
50
|
+
```
|
|
51
|
+
/agent-foundry:invoke critic "Review my authentication approach"
|
|
52
|
+
/agent-foundry:review
|
|
53
|
+
/agent-foundry:status
|
|
54
|
+
```
|
|
55
|
+
|
|
56
|
+
### Option 2: MCP Server Only
|
|
57
|
+
|
|
58
|
+
Install as an MCP server (no slash commands, just MCP tools):
|
|
38
59
|
|
|
39
60
|
```bash
|
|
40
61
|
npm install -g mcp-agent-foundry
|
|
41
62
|
```
|
|
42
63
|
|
|
43
|
-
|
|
64
|
+
Agent Foundry automatically registers with Claude Code during installation.
|
|
44
65
|
Restart Claude Code to activate.
|
|
45
66
|
|
|
46
67
|
To verify: Run `/mcp` in Claude Code - you should see `agent-foundry · ✓ connected`
|
|
@@ -59,6 +80,35 @@ This interactive wizard will:
|
|
|
59
80
|
|
|
60
81
|
---
|
|
61
82
|
|
|
83
|
+
## Plugin Commands
|
|
84
|
+
|
|
85
|
+
When installed as a Claude Code plugin, these slash commands are available:
|
|
86
|
+
|
|
87
|
+
| Command | Description | Example |
|
|
88
|
+
|---------|-------------|---------|
|
|
89
|
+
| `/agent-foundry:invoke` | Invoke an agent by role | `/agent-foundry:invoke critic "Review this plan"` |
|
|
90
|
+
| `/agent-foundry:review` | Get external code review | `/agent-foundry:review src/auth.ts` |
|
|
91
|
+
| `/agent-foundry:critique` | Get plan/architecture critique | `/agent-foundry:critique` |
|
|
92
|
+
| `/agent-foundry:compare` | Compare multiple agents | `/agent-foundry:compare coder,reviewer "Implement auth"` |
|
|
93
|
+
| `/agent-foundry:status` | Show provider health | `/agent-foundry:status` |
|
|
94
|
+
| `/agent-foundry:setup` | Configure providers | `/agent-foundry:setup` |
|
|
95
|
+
| `/agent-foundry:pipeline` | Run multi-agent pipeline | `/agent-foundry:pipeline` |
|
|
96
|
+
| `/agent-foundry:worktree` | Manage git worktrees | `/agent-foundry:worktree list` |
|
|
97
|
+
|
|
98
|
+
### Auto-Invoked Skills
|
|
99
|
+
|
|
100
|
+
The plugin also includes context-aware skills that Claude invokes automatically:
|
|
101
|
+
|
|
102
|
+
| Skill | Triggers When |
|
|
103
|
+
|-------|---------------|
|
|
104
|
+
| `second-opinion` | Making architectural decisions or choosing approaches |
|
|
105
|
+
| `code-review-suggest` | After significant code changes (50+ lines) |
|
|
106
|
+
| `failover-aware` | Provider errors occur (explains what happened) |
|
|
107
|
+
| `multi-agent-pipeline` | Complex multi-step tasks are described |
|
|
108
|
+
| `session-status` | Session starts (shows provider health) |
|
|
109
|
+
|
|
110
|
+
---
|
|
111
|
+
|
|
62
112
|
## How It Works
|
|
63
113
|
|
|
64
114
|
```
|