toolcapsule 0.1.0-alpha.11 → 0.1.0-alpha.12
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 +37 -12
- package/dist/cli.js +521 -283
- package/dist/cli.js.map +1 -1
- package/docs/agent-tool-compatibility.md +14 -3
- package/docs/architecture.md +1 -1
- package/docs/blog-launch.md +13 -10
- package/docs/concept.md +3 -3
- package/docs/importing-mcp.md +36 -7
- package/docs/launch.md +9 -8
- package/docs/next-steps.md +4 -3
- package/examples/feishu/README.md +4 -1
- package/llms.txt +9 -1
- package/package.json +2 -1
- package/skills/toolcapsule/SKILL.md +122 -0
package/README.md
CHANGED
|
@@ -52,18 +52,20 @@ This is the practical version of a lazy MCP workflow:
|
|
|
52
52
|
Already configured MCP in Claude Code, GitHub Copilot / VS Code, OpenCode, Gemini CLI, or Cursor? Import it instead of retyping URLs and commands:
|
|
53
53
|
|
|
54
54
|
```bash
|
|
55
|
-
tcap
|
|
56
|
-
tcap
|
|
55
|
+
tcap mcp list --include-user
|
|
56
|
+
tcap mcp enable docs --as feishu --target claude
|
|
57
57
|
```
|
|
58
58
|
|
|
59
59
|
ToolCapsule scans common workspace MCP config files such as `.mcp.json`, `.vscode/mcp.json`, `opencode.json`, `.gemini/settings.json`, and `.cursor/mcp.json`, then creates:
|
|
60
60
|
|
|
61
61
|
```text
|
|
62
|
-
|
|
62
|
+
~/.toolcapsule/profiles/<server>.json
|
|
63
63
|
.claude/skills/<server>-mcp/SKILL.md
|
|
64
64
|
```
|
|
65
65
|
|
|
66
|
-
Use `--target copilot`, `--target opencode`, `--target agents`, or `--target all` to write skills for other agents. User-level MCP configs are only inspected when you opt in with `--include-user`.
|
|
66
|
+
MCP profiles are user-level by default because MCP server connections are usually reused across projects. Use `--local` to store a profile in `.toolcapsule/profiles/` for the current workspace. Use `--target copilot`, `--target opencode`, `--target agents`, or `--target all` to write skills for other agents. User-level MCP configs are only inspected when you opt in with `--include-user`.
|
|
67
|
+
|
|
68
|
+
Imports link back to the original MCP config by default so ToolCapsule does not duplicate private MCP URLs. Use `--copy` when you want a stable ToolCapsule snapshot profile.
|
|
67
69
|
|
|
68
70
|
## What is a ToolCapsule?
|
|
69
71
|
|
|
@@ -77,12 +79,30 @@ Instead of making the model hold everything in the prompt, ToolCapsule stores he
|
|
|
77
79
|
|
|
78
80
|
## Quick start
|
|
79
81
|
|
|
82
|
+
AI-first onboarding: install the ToolCapsule Skill into your coding agent, then ask the agent to inventory MCPs and capsule the heavy ones.
|
|
83
|
+
|
|
84
|
+
```bash
|
|
85
|
+
npx skills add RainSunMe/toolcapsule --skill toolcapsule
|
|
86
|
+
```
|
|
87
|
+
|
|
88
|
+
For a specific agent or global install:
|
|
89
|
+
|
|
90
|
+
```bash
|
|
91
|
+
npx skills add RainSunMe/toolcapsule --skill toolcapsule --agent claude-code
|
|
92
|
+
npx skills add RainSunMe/toolcapsule --skill toolcapsule --global --agent claude-code
|
|
93
|
+
```
|
|
94
|
+
|
|
95
|
+
Then tell your agent:
|
|
96
|
+
|
|
97
|
+
> Use the ToolCapsule Skill. Scan my MCP servers, pick the heavy ones, and turn them into lazy ToolCapsule Skills.
|
|
98
|
+
|
|
99
|
+
Manual CLI workflow:
|
|
100
|
+
|
|
80
101
|
```bash
|
|
81
102
|
npm i -g toolcapsule
|
|
82
103
|
|
|
83
|
-
toolcapsule
|
|
84
|
-
toolcapsule
|
|
85
|
-
toolcapsule import --name feishu --target claude
|
|
104
|
+
toolcapsule mcp list --include-user
|
|
105
|
+
toolcapsule mcp enable docs --as feishu --target claude
|
|
86
106
|
toolcapsule tools feishu --brief
|
|
87
107
|
toolcapsule schema feishu create-doc
|
|
88
108
|
toolcapsule call feishu create-doc @args.json --save-run
|
|
@@ -101,12 +121,13 @@ tcap call feishu create-doc @args.json --save-run
|
|
|
101
121
|
```text
|
|
102
122
|
.claude/skills/feishu-mcp/
|
|
103
123
|
SKILL.md # lightweight Agent Skill entrypoint
|
|
104
|
-
toolcapsule.config.json # MCP transport/profile config
|
|
105
124
|
scripts/README.md
|
|
106
125
|
|
|
126
|
+
~/.toolcapsule/
|
|
127
|
+
profiles/feishu.json # user-level MCP profile, reusable across projects
|
|
128
|
+
|
|
107
129
|
.toolcapsule/
|
|
108
|
-
|
|
109
|
-
runs/feishu/ # auditable tool call records
|
|
130
|
+
runs/feishu/ # workspace-local auditable tool call records
|
|
110
131
|
```
|
|
111
132
|
|
|
112
133
|
This lets your agent discover the Skill instead of carrying the full MCP schema in every turn.
|
|
@@ -153,9 +174,13 @@ Results vary by MCP server, model, and host.
|
|
|
153
174
|
```text
|
|
154
175
|
toolcapsule init <name> --url <remote-mcp-url> --target claude
|
|
155
176
|
toolcapsule init <name> --command <stdio-command> --arg <arg> --target claude
|
|
156
|
-
toolcapsule
|
|
177
|
+
npx skills add RainSunMe/toolcapsule --skill toolcapsule
|
|
178
|
+
toolcapsule mcp list --include-user
|
|
179
|
+
toolcapsule mcp enable <server> --as <profile> --target claude
|
|
180
|
+
toolcapsule mcp disable <profile>
|
|
181
|
+
toolcapsule mcp disable <server> --native
|
|
157
182
|
toolcapsule import --dry-run
|
|
158
|
-
toolcapsule import --name <server> --target claude
|
|
183
|
+
toolcapsule import --name <server> --as <profile> --target claude
|
|
159
184
|
toolcapsule import --all --target all
|
|
160
185
|
toolcapsule tools <profile> --brief
|
|
161
186
|
toolcapsule describe <profile> <tool> --brief
|