mukunda1729 commited on
Commit
5219695
·
verified ·
1 Parent(s): 259f98f

Initial: 22 MCP tool-call fixtures (valid + invalid)

Browse files
Files changed (2) hide show
  1. README.md +71 -0
  2. data.jsonl +22 -0
README.md ADDED
@@ -0,0 +1,71 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ license: mit
3
+ language:
4
+ - en
5
+ tags:
6
+ - mcp
7
+ - model-context-protocol
8
+ - tool-use
9
+ - testing
10
+ - agents
11
+ - validation
12
+ size_categories:
13
+ - n<1K
14
+ configs:
15
+ - config_name: default
16
+ data_files:
17
+ - split: train
18
+ path: data.jsonl
19
+ ---
20
+
21
+ # mcp-tool-test-fixtures
22
+
23
+ 22 test fixtures for common MCP (Model Context Protocol) tool calls — a mix of valid and invalid args across 8 categories. Use these to test tool-arg validators, sandboxes, or any layer that sits between an LLM and a tool.
24
+
25
+ ## Categories
26
+
27
+ | Category | Count |
28
+ |---|---|
29
+ | `filesystem` | 5 |
30
+ | `git` | 3 |
31
+ | `github` | 3 |
32
+ | `slack` | 2 |
33
+ | `http` | 3 |
34
+ | `shell` | 2 |
35
+ | `calculator` | 2 |
36
+ | `web` | 2 |
37
+
38
+ About **half** are intentionally invalid — missing required fields, dangerous commands, escape attempts, wrong types — so your validator gets exercised on negative cases too.
39
+
40
+ ## Schema
41
+
42
+ ```jsonc
43
+ {
44
+ "id": "string",
45
+ "tool": "string", // dotted MCP tool name
46
+ "args": { ... }, // args the LLM produced
47
+ "expected": {
48
+ "valid": true | false,
49
+ "reason": "string" // present when invalid
50
+ },
51
+ "category": "string"
52
+ }
53
+ ```
54
+
55
+ ## Quickstart
56
+
57
+ ```python
58
+ from datasets import load_dataset
59
+ ds = load_dataset("mukunda1729/mcp-tool-test-fixtures", split="train")
60
+ invalid = [r for r in ds if not r["expected"]["valid"]]
61
+ print(f"{len(invalid)} invalid cases for negative testing")
62
+ ```
63
+
64
+ ## Related
65
+
66
+ - [`agentvet` on PyPI](https://pypi.org/project/agentvet-py/) — wraps tool calls, validates args, returns LLM-friendly retry hints
67
+ - [The Agent Reliability Stack](https://mukundakatta.github.io/agent-stack/)
68
+
69
+ ## License
70
+
71
+ MIT.
data.jsonl ADDED
@@ -0,0 +1,22 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {"id": "fs-read-1", "tool": "filesystem.read_file", "args": {"path": "/tmp/example.txt"}, "expected": {"valid": true}, "category": "filesystem"}
2
+ {"id": "fs-read-2", "tool": "filesystem.read_file", "args": {"path": "../../../etc/passwd"}, "expected": {"valid": false, "reason": "path_escape"}, "category": "filesystem"}
3
+ {"id": "fs-read-3", "tool": "filesystem.read_file", "args": {}, "expected": {"valid": false, "reason": "missing_required: path"}, "category": "filesystem"}
4
+ {"id": "fs-write-1", "tool": "filesystem.write_file", "args": {"path": "/tmp/out.txt", "content": "hello"}, "expected": {"valid": true}, "category": "filesystem"}
5
+ {"id": "fs-write-2", "tool": "filesystem.write_file", "args": {"path": "/tmp/out.txt", "content": 123}, "expected": {"valid": false, "reason": "wrong_type: content must be string"}, "category": "filesystem"}
6
+ {"id": "git-status-1", "tool": "git.status", "args": {"repo": "/Users/me/proj"}, "expected": {"valid": true}, "category": "git"}
7
+ {"id": "git-commit-1", "tool": "git.commit", "args": {"repo": "/Users/me/proj", "message": "fix bug"}, "expected": {"valid": true}, "category": "git"}
8
+ {"id": "git-commit-2", "tool": "git.commit", "args": {"repo": "/Users/me/proj"}, "expected": {"valid": false, "reason": "missing_required: message"}, "category": "git"}
9
+ {"id": "github-search-1", "tool": "github.search_issues", "args": {"q": "is:open label:bug"}, "expected": {"valid": true}, "category": "github"}
10
+ {"id": "github-issue-1", "tool": "github.create_issue", "args": {"repo": "owner/name", "title": "bug", "body": "details"}, "expected": {"valid": true}, "category": "github"}
11
+ {"id": "github-issue-2", "tool": "github.create_issue", "args": {"repo": "owner/name"}, "expected": {"valid": false, "reason": "missing_required: title"}, "category": "github"}
12
+ {"id": "slack-send-1", "tool": "slack.send_message", "args": {"channel": "#general", "text": "hi"}, "expected": {"valid": true}, "category": "slack"}
13
+ {"id": "slack-send-2", "tool": "slack.send_message", "args": {"channel": "#general"}, "expected": {"valid": false, "reason": "missing_required: text"}, "category": "slack"}
14
+ {"id": "fetch-1", "tool": "fetch.get", "args": {"url": "https://api.example.com/v1/data"}, "expected": {"valid": true}, "category": "http"}
15
+ {"id": "fetch-2", "tool": "fetch.get", "args": {"url": "file:///etc/passwd"}, "expected": {"valid": false, "reason": "scheme_not_allowed"}, "category": "http"}
16
+ {"id": "fetch-3", "tool": "fetch.get", "args": {"url": "not-a-url"}, "expected": {"valid": false, "reason": "invalid_url"}, "category": "http"}
17
+ {"id": "shell-1", "tool": "shell.exec", "args": {"cmd": "ls"}, "expected": {"valid": true}, "category": "shell"}
18
+ {"id": "shell-2", "tool": "shell.exec", "args": {"cmd": "rm -rf /"}, "expected": {"valid": false, "reason": "dangerous_command"}, "category": "shell"}
19
+ {"id": "calc-1", "tool": "calculator.eval", "args": {"expr": "2+2"}, "expected": {"valid": true}, "category": "calculator"}
20
+ {"id": "calc-2", "tool": "calculator.eval", "args": {"expr": "__import__('os').system('ls')"}, "expected": {"valid": false, "reason": "unsafe_expression"}, "category": "calculator"}
21
+ {"id": "search-1", "tool": "web.search", "args": {"q": "python tutorials", "num": 5}, "expected": {"valid": true}, "category": "web"}
22
+ {"id": "search-2", "tool": "web.search", "args": {"q": "", "num": 5}, "expected": {"valid": false, "reason": "empty_query"}, "category": "web"}