MetaGO Algorithms

io.github.metago-aiPublished by GitHub account metago-aiv1.0.2Updated Nov 3, 2026

927 deterministic algorithms as 57 MCP tools: coupling, value, distance, correlation, reduction.

VerifiedSTDIODesktop onlyDeveloper ToolsData & AnalyticsAI & ML

Overview

AI-generated overview

Local MCP server exposing 57 tools backed by 927 deterministic algorithms for coupling, value scoring, distance, correlation and dimensionality reduction.

What you can do
  • Execute any named algorithm with parameters, or run it synchronously, and read server statistics
  • List, search and inspect algorithms and their families, and get recommendations for a task
  • Compute coupling scores, matrices, clustering, trends and co-occurrence frequencies
  • Score multi-dimensional value and build value vectors
  • Measure distances and similarity (Euclidean, Manhattan, Chebyshev, cosine, Dice, Jaccard) and correlation (Pearson, Spearman, Kendall, covariance)
  • Reduce dimensions with PCA and SVD, run security and bias checks, and process frequency and time data
When to use it
Useful when an assistant needs reproducible numeric analysis rather than free-form reasoning: similarity and distance between vectors, correlation statistics, feature reduction, coupling or value scoring, and quick lookup of which algorithm fits a task.
Requirements
Runs as a local process over stdio; install the npm package @metago-ai/algorithms and start it with node or npx. No authentication, environment variables or headers are declared. The README states it depends on @metago-ai/[email protected].
Before you install
The README says every tool is driven by an external engine with memory, evolution and decision-lock components, and that it can generate new algorithms and skill files; review what that engine stores or writes locally before use. Several tools are named for memory, learning, evolution and decision operations, so check whether they persist state. No credentials are requested.

Installation

In SourceWeft

  1. Open MetaGO Algorithms in the dashboard and add it to a workspace.
  2. Enable the server for the chats that should use its tools.

Desktop only via STDIO. STDIO servers start a local process, so they need the SourceWeft desktop host.

Other MCP clients

Follow the launch instructions in the repository.

README

@metago-ai/algorithms — 927 算法 MCP 服务器

MetaGO Intelligent Lifeform 的算法核心 — 57 工具 / 927 算法 / 14 触发大类 / 4 层架构，Engine V2.1.0 硬驱动。

[npm] [License: MIT] [Tools] [Algorithms] [Engine]

What it is

@metago-ai/algorithms is the algorithmic backbone of the MetaGO Intelligent Lifeform. It exposes 57 MCP tools backed by 927 algorithms across 14 trigger categories, organized in a 4-layer architecture. Every tool is hard-driven by @metago-ai/[email protected] (KMWI memory + evolution engine + skill generator).

This is not a library of math functions. It is a Model Context Protocol (MCP) server that gives any MCP-compatible AI agent (Trae, Claude Code, Codex, Cursor, etc.) real-time access to coupling calculation, value assessment, distance metrics, correlation analysis, dimensionality reduction, security audit, frequency processing, and 7 more trigger categories.

Architecture — 4 layers / 14 trigger categories

LayerNameToolsAlgorithmsTrigger Categories
L1通用入口层 (General Entry)393
L2检索发现层 (Retrieval & Discovery)4184
L3高频专用层 (High-Frequency Specialized)305404
L4族级工具层 (Family Tools)203603
Total—5792714

The 14 trigger categories

#CategoryLayerTrigger condition
1Algorithm executionL1Agent needs to execute an algorithm
2Algorithm retrievalL1Agent needs to find an algorithm by name/family/capability
3StatisticsL1Agent needs server statistics
4Family listingL2Agent needs to list algorithm families
5Algorithm searchL2Agent needs semantic search over algorithms
6Info lookupL2Agent needs detailed info on a specific algorithm
7Proactive suggestionL2Agent needs algorithm recommendations for a task
8Coupling calculationL3Task involves relationship strength / coupling degree
9Value assessmentL3Task involves multi-dimensional value scoring
10Distance metricsL3Task involves similarity / distance between vectors
11Correlation analysisL3Task involves statistical correlation
12Dimensionality reductionL3Task involves PCA / SVD / feature reduction
13Security auditL3Task involves vulnerability / security assessment
14Frequency processingL3Task involves frequency-domain analysis

Install

bash
npm install @metago-ai/algorithms

Use as MCP server

Trae / Claude Code / Cursor / any MCP client

Add to your MCP client config:

json
{  "mcpServers": {    "metago-algorithms": {      "command": "node",      "args": ["node_modules/@metago-ai/algorithms/index.js"]    }  }}

Standalone

bash
npx @metago-ai/algorithms# ornode node_modules/@metago-ai/algorithms/index.js

Tool list (57 tools)

L1 — General Entry (3 tools)

  • execute — Execute any algorithm by name with parameters
  • execute_sync — Synchronous algorithm execution
  • get_statistics — Server statistics (tool count, algorithm count, layer distribution)

L2 — Retrieval & Discovery (4 tools)

  • list_algorithms — List all 927 algorithms with metadata
  • list_families — List 14 algorithm families
  • search_algorithms — Semantic search over algorithm descriptions
  • get_algorithm_info — Detailed info on a specific algorithm

L3 — High-Frequency Specialized (30 tools)

Coupling (8 tools)
  • coupling_calculate · coupling_clustering · coupling_trend · normalize_coupling
  • build_coupling_matrix · sort_coupling_scores · is_superconductive · cooccurrence_frequency
Value Assessment (3 tools)
  • value_assess · build_value_vector · evaluate_bidirectional
Distance Metrics (6 tools)
  • euclidean_distance · manhattan_distance · chebyshev_distance
  • cosine_similarity · dice_coefficient · jaccard_coefficient
Correlation (4 tools)
  • pearson_correlation · spearman_correlation · kendall_tau · covariance
Dimensionality Reduction (3 tools)
  • extract_pca · simplified_svd · reduce_dimension
Security & Audit (3 tools)
  • security_assess · audit_check · bias_detect
Frequency & Time (3 tools)
  • frequency_process · time_analyze · time_decay_cooccurrence

L4 — Family Tools (20 tools)

Fuzzy & Semantic (3 tools)
  • fuzzy_string_match · semantic_similarity · weighted_cosine
Reasoning (3 tools)
  • reasoning_chain · logic_reason · conflict_resolve
Asymmetry & Symmetry (3 tools)
  • detect_asymmetry · record_symmetric · identify_strong_pairs · identify_weak_pairs
Negentropy (1 tool)
  • negentropy_measure
Intuition (1 tool)
  • intuition_gauge
Lifeform Operations (8 tools)
  • memory_operate · learning_learn · creation_create · evolution_evolve
  • decision_decide · dialog_manage · idea_generate · proactive_suggest

Engine dependency

This package is hard-driven by @metago-ai/[email protected]:

  • KMWI Memory — 4-layer memory (Knowledge → Memory → Wisdom → Intuition) backs algorithm state
  • EvolutionEngine — 5-stage meta-evolution can generate new algorithms on demand
  • DecisionLock — 4-gate verification before every algorithm output
  • SkillGenerator — generates new algorithm SKILL.md files from internal patterns

License

MIT © 2026 易霄 / 元构光年（成都）人工智能科技有限公司

Links

Source: packages/algorithms/README.md at commit de8479e

Tools

0
SourceWeft has not started this package in its sandbox yet, so its tools are not known.

Version history

1
  1. v1.0.2LatestSep 27, 2026

Score

50% of the applicable checks met

How complete this listing is, computed from the registry entry, the endpoint and public sources. Not a rating of the server and not a statement that it is safe or working today: reachability is a dated check, and no installs or reviews go into the score.

  • Reached in the last 30 days25 pts

    Keep the endpoint answering the MCP handshake; SourceWeft probes it every 30 days.

  • Auth declared10 pts

    Declare the credential headers in the registry entry, or keep the endpoint answering the probe so its auth can be classified.

  • Tools listed15 pts

    Answer tools/list without a credential, or ship a package that starts in a sandbox without one.

  • DescriptionRequired15 pts
  • Logo10 pts
  • Documentation15 pts
  • License10 pts