Skip to main content
Glama

Agent Vending Factory

by io.github.RubenTay

tool_compute_sandbox

POST /tools/tool_compute_sandbox/run — Executes Python 3.12 code in an isolated subprocess with a 5-second hard timeout. Input: {python_code: string, input_data: any (optional, bound as variable 'input_data')}. Output: {success, result, stdout (capped 50KB), execution_time_ms, error_type}. Return value: assign to 'result' variable. Pre-loaded: math, json, re, statistics, itertools, functools, collections, decimal, datetime, random, hashlib, base64. Blocked: import, open(), eval(), exec(), os, sys, network, class definitions, dunder attributes. error_type values: syntax_error | security_error | runtime_error | timeout_error. Cost: $0.1500 USDC per call.

Input Schema

TableJSON Schema
NameRequiredDescriptionDefault
input_dataNoJSON-serializable value (dict, list, str, number, bool, or null) passed as the variable 'input_data' inside the sandbox. Omit or pass null if the code has no external input.
python_codeYesPython 3.12 source code to execute. No import statements. Set 'result = <value>' to return a value. Pre-loaded modules are in scope: math, json, re, statistics, itertools, functools, collections, decimal, datetime, random, string, textwrap, hashlib, base64, struct, copy, pprint.

Schema Changelog

Changes observed during successful MCP inspections. Dates show when Glama detected each change.

  1. First observed

TDQS

A4.5/5.0
Behavior5/5

Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?

With no annotations provided, the description carries the full burden and excels: it discloses the isolated subprocess, hard timeout, curated pre-loaded modules, blocked operations (import, open, eval, exec, os, sys, network, classes, dunders), stdout cap, error_type enum values, and per-call cost. This is exceptionally transparent about behavior and safety boundaries.

Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.

Conciseness5/5

Is the description appropriately sized, front-loaded, and free of redundancy?

The description is dense but well-structured with labeled sections (Input, Output, Pre-loaded, Blocked, error_type, Cost). Every sentence carries critical information, and the format makes it easy to scan. It packs a lot of details without redundancy.

Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.

Completeness5/5

Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?

For a code-execution tool with no output schema, the description is remarkably complete. It covers input format, output shape, error types, timeout, resource limits, allowed modules, forbidden operations, and cost. An agent has all the information needed to decide whether to use the tool and how to interpret its results.

Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.

Parameters4/5

Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?

Schema coverage is 100%, so the baseline is 3. The description adds meaningful semantics beyond the schema by explaining that input_data is bound to a variable named 'input_data', that the return value must be assigned to 'result', and by enumerating the output fields (success, result, stdout, execution_time_ms, error_type). This extra context helps the agent use the parameters correctly.

Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.

Purpose5/5

Does the description clearly state what the tool does and how it differs from similar tools?

The description opens with a specific verb ('Executes') and a precise resource ('Python 3.12 code in an isolated subprocess'), plus a 5-second timeout. This clearly differentiates the tool from siblings like tool_data_transformer and tool_zar_prices, which are not code-execution tools.

Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.

Usage Guidelines3/5

Does the description explain when to use this tool, when not to, or what alternatives exist?

The description implies usage by detailing execution constraints, blocked operations, and cost, but it never explicitly states when to use this tool versus alternatives. There is no mention of 'use this when you need to run arbitrary Python' or any exclusions. The purpose is clear, but usage guidance is only implicitly conveyed.

Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.

All Agent Vending Factory toolsInspect
Try in Browser
RayrunBuild, deploy, and operate MCP servers without becoming an infrastructure team.ray.runAd

Glama MCP Gateway

Add one secure layer between your agents and this server.

TDQS

A4.4/5.0
Disambiguation5/5

Each tool covers a completely different domain: Python execution, raw-text JSON extraction, South African airport flights, and ZAR crypto prices. There is no overlap in purpose, input shape, or use case, so agents should not confuse them.

Naming Consistency4/5

All tools share a 'tool_' prefix and lower_snake_case formatting, giving the set a recognizable pattern. The suffix style is mostly consistent, though 'compute_sandbox' and 'data_transformer' are generic capability nouns while 'sa_airport_oracle' and 'zar_prices' are more domain-branded, and none uses a verb-action pattern.

Tool Count4/5

Four tools is a manageable, non-bloated count for a utility/vending server. It is slightly lean given the broad 'Agent Vending Factory' name, but each tool is substantial and earns its place.

Completeness3/5

Each tool covers its own narrow operation well, but the set lacks a unifying domain or shared workflow; the compute, transformer, flight, and pricing endpoints are isolated one-shots. There is also no meta-tool for discovering, chaining, or managing the vended capabilities, so the surface feels like a grab-bag rather than a complete product.

Resources