@tankpkg/mcp-server 0.0.0-nightly.20260401.49863f7

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 ADDED
@@ -0,0 +1,147 @@
1
+ # @tankpkg/mcp-server
2
+
3
+ MCP (Model Context Protocol) server for Tank — manage AI agent skills directly from your editor. Full CLI parity: every `tank` command is available as an MCP tool.
4
+
5
+ ## Features
6
+
7
+ ### Authentication
8
+
9
+ - **login** — Authenticate with Tank via GitHub OAuth
10
+ - **logout** — Clear stored credentials
11
+ - **whoami** — Show current user info
12
+
13
+ ### Project Setup
14
+
15
+ - **init-skill** — Create `skills.json` and `SKILL.md` scaffold
16
+
17
+ ### Publishing & Discovery
18
+
19
+ - **publish-skill** — Publish a skill to the Tank registry (with dry-run support)
20
+ - **search-skills** — Search the Tank registry for skills
21
+ - **skill-info** — Get detailed information about a specific skill
22
+
23
+ ### Installation & Management
24
+
25
+ - **install-skill** — Install a skill with SHA-512 verification
26
+ - **update-skill** — Update skills within semver range
27
+ - **remove-skill** — Remove a skill and clean up lockfile
28
+
29
+ ### Security & Verification
30
+
31
+ - **scan-skill** — Scan any directory for security issues (skills.json not required)
32
+ - **verify-skills** — Verify lockfile integrity
33
+ - **audit-skill** — Show security scan results and verdict
34
+ - **skill-permissions** — Display per-skill permission summary
35
+
36
+ ### Agent Integration
37
+
38
+ - **link-skill** — Symlink a skill into an agent workspace
39
+ - **unlink-skill** — Remove a skill symlink
40
+
41
+ ### Diagnostics
42
+
43
+ - **doctor** — Check config, auth, registry connectivity, and Node.js version
44
+
45
+ ## Installation
46
+
47
+ ### Claude Code
48
+
49
+ Add to `.claude/settings.json` or your project's `.mcp.json`:
50
+
51
+ ```json
52
+ {
53
+ "mcpServers": {
54
+ "tank": {
55
+ "command": "npx",
56
+ "args": ["-y", "@tankpkg/mcp-server"]
57
+ }
58
+ }
59
+ }
60
+ ```
61
+
62
+ ### Cursor
63
+
64
+ Add to `~/.cursor/mcp.json`:
65
+
66
+ ```json
67
+ {
68
+ "mcpServers": {
69
+ "tank": {
70
+ "command": "npx",
71
+ "args": ["-y", "@tankpkg/mcp-server"]
72
+ }
73
+ }
74
+ }
75
+ ```
76
+
77
+ ### VS Code (Copilot)
78
+
79
+ Add to `.vscode/mcp.json`:
80
+
81
+ ```json
82
+ {
83
+ "servers": {
84
+ "tank": {
85
+ "command": "npx",
86
+ "args": ["-y", "@tankpkg/mcp-server"]
87
+ }
88
+ }
89
+ }
90
+ ```
91
+
92
+ ## Authentication
93
+
94
+ The MCP server shares authentication with the Tank CLI. If you've already run `tank login`, you're authenticated!
95
+
96
+ Alternatively, set the `TANK_TOKEN` environment variable:
97
+
98
+ ```json
99
+ {
100
+ "mcpServers": {
101
+ "tank": {
102
+ "command": "npx",
103
+ "args": ["-y", "@tankpkg/mcp-server"],
104
+ "env": {
105
+ "TANK_TOKEN": "tank_your_token_here"
106
+ }
107
+ }
108
+ }
109
+ }
110
+ ```
111
+
112
+ ## Usage Examples
113
+
114
+ Once configured, talk to your AI agent naturally:
115
+
116
+ ```
117
+ "Initialize a new skill in this directory"
118
+ "Scan my skill for security issues"
119
+ "Publish my skill to Tank"
120
+ "Install @acme/code-review"
121
+ "Update all my skills"
122
+ "Search Tank for testing skills"
123
+ "Show permissions for my installed skills"
124
+ "Run tank doctor to check my setup"
125
+ "Audit @acme/code-review for security issues"
126
+ "Link this skill to my Claude workspace"
127
+ ```
128
+
129
+ ## Development
130
+
131
+ ```bash
132
+ # Install dependencies
133
+ bun install
134
+
135
+ # Build
136
+ bun build
137
+
138
+ # Run tests
139
+ bun test
140
+
141
+ # Start the server (stdio mode)
142
+ bun start
143
+ ```
144
+
145
+ ## License
146
+
147
+ MIT
@@ -0,0 +1 @@
1
+ export { };