Skip to main content

Skilly

Skilly Logo

Autonomous, 100% LLM-Free Codebase Capability Analyzer & Interactive Knowledge Graph Synthesizer

PyPI Package Author Arastu Thakur GitHub arastuthakur LinkedIn arastuthakur

Zero LLM Air Gapped Architecture Grade Polyglot Zero DB License


Overview

Skilly is a portable, enterprise-grade static analysis engine and capability catalog synthesizer designed to run in any codebase, on any operating system, in under 1 second — with zero external LLM dependencies, zero API tokens, and zero external databases.

Point Skilly at any project directory (skilly . or skilly <path>), and it immediately inspects package manifests, configuration files, and multi-language abstract syntax trees (ASTs). It synthesizes:

  1. skills.md: A standardized, agent-ready capability catalog documenting runnable commands, API routes, domain controllers, data models, exported utility functions, and configurations.
  2. knowledge_graph.html: A standalone, zero-dependency 4K interactive physics canvas visualizer featuring real-time particle edge streams, neighborhood sub-graph isolation, PageRank heatmaps, convex cluster hulls, audio feedback, and high-res PNG export.
  3. knowledge_graph.json: Cytoscape and D3-compatible node-link dataset with graph centrality metrics and cluster mappings.
  4. knowledge_graph.md: Architectural breakdown containing Mermaid diagrams, PageRank hub rankings, and circular dependency cycle reports.

Empirical Benchmark: Skilly vs. Infigraph vs. Graphifyy

Measured directly by executing all three tools on the same codebase.

Dimension / Benchmark Metric Skilly (by Arastu Thakur) Infigraph (by Intuit) Graphifyy
Measured Runtime (Sample Project) 0.14s (Fastest) 1.20s 8.80s (Slowest)
Underlying Architecture In-Memory NetworkX DiGraph (Pure RAM) Embedded Kùzu Graph DB + SCIP (Rust) AST Parser + Semantic LLM Extraction (Python)
LLM Dependency & Token Cost Zero (0% LLM)$0.00 Forever Zero (0% LLM)$0.00 Forever Requires LLM for docs, semantics & naming
Database & Disk Footprint 0 MB (Pure in-memory RAM execution) Disk DB (.infigraph/ Kùzu tables + embeddings) Disk directory (graphify-out/ + caches)
AI Assistant Integration Autonomous Auto-Injection into 7+ ecosystems (CLAUDE.md, .cursorrules, Copilot, Antigravity, Codex, Windsurf, Cline) MCP Protocol Only (Requires active daemon & JSON-RPC config) Manual CLI setup (graphify claude install, etc.)
Generated Capability Catalog Native skills.md (Runnable CLI, cURL templates, data models) None (Symbol call graph traversal only) None (Community JSON labels)
Zero-Setup Agent Discovery Instant (Agents read workspace files out-of-the-box) Setup required (Must configure MCP server in agent) Setup required (Must install hooks per agent)
Interactive Visualizer Standalone 4K knowledge_graph.html (Particles, hulls, audio, PNG export — zero server needed) Web UI (Requires active local HTTP daemon running) Collapsible D3 Tree (Requires separate CLI command)
CI/CD Quality Gates Native (--fail-on-grade=A, --fail-on-cycles, --json-summary) PR review blast radius & affected test detection None
Polyglot Coverage 50+ Languages + 16 Manifest & build formats 62 Languages via Tree-Sitter, ANTLR, SCIP Code files (Python, JS, TS, etc.)
Deterministic Code Truth 100% Reproducible (Exact syntax trees & hashes) 100% Reproducible (AST & compiler indexers) Non-deterministic (Subject to LLM variations)

Interactive Visualizer Architecture (knowledge_graph.html)

The interactive knowledge graph visualizer generated by Skilly is 100% self-contained (zero external CDN scripts required to render) and includes:

  • Live Particle Streams: Flowing neon energy pulses traveling along directed dependency edges to highlight active architectural data flow.
  • Neighborhood Sub-Graph Isolation: Click any file, function, class, or route to isolate its 1-hop or 2-hop dependency neighborhood while dimming unrelated nodes.
  • Architectural Convex Hulls: Chromatic glowing clusters grouping related subsystems and domains together.
  • PageRank Centrality Heatmap: Instant toggle between Categorical Type coloring and PageRank Centrality Heatmap (cyan -> amber -> neon rose) to expose architectural bottlenecks.
  • Integrated Minimap Radar: Live navigational overview with real-time camera viewport tracking.
  • 4K PNG Snapshot Export: One-click rasterization of the canvas at full display resolution with dark-mode contrast.
  • Synthesizer Haptics: Subtle Web Audio synthesizer chimes providing interactive auditory feedback on node hover and selection.

Synthesized Artifact Specifications

When you run skilly <project>, Skilly synthesizes four production-grade artifacts in your project directory:

my-project/
├── skills.md              # Complete capability catalog (commands, APIs, services, models)
├── knowledge_graph.html   # Standalone 4K interactive physics canvas graph visualizer
├── knowledge_graph.json   # Cytoscape/D3-compatible graph node-link dataset
└── knowledge_graph.md     # Architectural report with Mermaid diagrams & PageRank hubs

1. skills.md

Engineered for both human engineers and AI coding agents (such as Claude, Antigravity, Cursor, and Copilot). Contains:

  • Architecture Health Grade: Letter grade (A+ to D) and score.
  • Runnable CLI Commands & Workflows: Extracted from package.json, Makefile, Dockerfile, and CLI decorators (Click, Typer, Argparse). Includes exact execution cheat-sheets.
  • API Endpoints: Full HTTP method, route, parameters, and generated curl invocation templates for FastAPI, Express, Next.js, Flask, Gin, and Spring Boot.
  • Domain Services & Models: Extracted classes, methods, Pydantic schemas, TypeScript interfaces, and Go/Rust structs.
  • Environment & Configuration: Documented environment variables with default values and source file locations.
  • Architectural Hubs: Top PageRank-central components that anchor the codebase.

2. knowledge_graph.html

An interactive force-directed visualizer that can be opened in any web browser without running a server (file:// compatible) or served via skilly --serve.

3. knowledge_graph.md

A GitHub-flavored markdown report featuring:

  • Mermaid dependency diagrams.
  • Top architectural hubs ranked by PageRank importance.
  • Strongly connected components (circular reference cycles).
  • Domain cluster breakdown.

Architecture Health & CI/CD Quality Gates

Skilly computes a holistic structural health score for your codebase:

$$\text{Health Score} = 100 - (\text{Cycle Penalty}) + (\text{Doc Coverage Bonus}) + (\text{Modularity Ratio})$$

  • Circular Reference Detection: Identifies toxic circular dependencies (A -> B -> C -> A) using Tarjan's Strongly Connected Components algorithm.
  • Documentation Density: Quantifies docstring and documentation coverage across public functions and classes.
  • Modularity Ratio: Evaluates inter-cluster vs intra-cluster coupling.
  • Letter Grade: Awards a grade (A+, A, B, C, D).

CI/CD Quality Gates

Integrate Skilly directly into your continuous integration pipeline to block regressions:

# Fail build if circular dependencies exist:
skilly . --fail-on-cycles

# Fail build if architecture health falls below required grade:
skilly . --fail-on-grade=A

# Output structured JSON for pipeline integration:
skilly . --json-summary

Autonomous AI Assistant Context Injection

Whenever Skilly runs, it autonomously injects project capabilities, executable workflows, and architectural topology into leading AI coding environments:

AI Assistant / Tool Injected Configuration Target Purpose & Directives
Claude & Claude Code CLAUDE.md Primary instruction manual for Claude Code CLI and Anthropic projects
GitHub Copilot .github/copilot-instructions.md Contextual instruction file read by Copilot Chat & Agent mode
Cursor IDE .cursorrules & .cursor/rules/skilly.mdc Universal rules + MDC glob rules instructing Cursor on repository skills
Google Antigravity AGENTS.md, GEMINI.md, & .agents/rules/skilly.md Global agent directives and workspace rules for Antigravity coding agents
OpenAI Codex & ChatGPT CODEX.md System context prompt and task capabilities for OpenAI Codex runners
Windsurf (Codeium) .windsurfrules Cascade agent rules linking to runnable commands and data models
Cline & Roo Code .clinerules Autonomous task rules guiding Cline through project APIs and hubs

Non-Destructive & Idempotent

Skilly encapsulates its injected guidance within safe marker comments:

<!-- SKILLY_INJECTION_START -->
... [Autonomous Context, Skills References, and Architectural Health] ...
<!-- SKILLY_INJECTION_END -->

If you already have custom prompts in your CLAUDE.md or .cursorrules, Skilly preserves all existing instructions and non-destructively refreshes only the Skilly block. Multiple runs update in place without duplicating text.

CLI Auto-Injection Control

# Enabled by default:
skilly .

# Target only specific assistants:
skilly . --ai-targets=claude,cursor,antigravity

# Disable AI injection entirely:
skilly . --no-inject-ai

Installation & Multi-Environment Support

Skilly is engineered to run seamlessly across development environments:

1. Python Package (PyPI)

# Install directly from PyPI (recommended):
pip install skilly-ai

# Or install from source:
pip install -e .

# Or install directly via Git:
pip install git+https://github.com/arastuthakur/skilly.git

2. Node.js / NPX Runner (JavaScript / TypeScript Environments)

# Direct execution:
node bin/skilly-node.js .

# Or global install:
npm install -g .
skilly .

3. Shell Executables & Installers

  • Linux & macOS: Run ./install.sh or ./bin/skilly .
  • Windows PowerShell: Run .\install.ps1 or .\bin\skilly.ps1 .
  • Windows Command Prompt: Run .\skilly.cmd .

4. Docker Container (Zero-Install)

docker run --rm -v $(pwd):/project ghcr.io/arastuthakur/skilly /project

5. GitHub Actions CI/CD Workflow

Add .github/workflows/skilly.yml to your repository to autonomously generate and commit updated capability catalogs and visualizers on every pull request.


Polyglot Language Coverage (50+ Languages)

Skilly's universal AST extraction engine supports over 50 languages via Tree-Sitter grammars and native AST parsers:

Family Languages & Grammars
Systems & Native C, C++, Rust, Go, Zig, Nim, D, Assembly
Managed & Enterprise Java, Kotlin, Scala, C#, F#, Swift, Dart (Flutter)
Web & Scripting Python, JavaScript, TypeScript, Ruby, PHP, Lua, Julia, Shell (Bash/Zsh), PowerShell
Functional Elixir, Erlang, Clojure, Haskell, OCaml, R
Schemas & Smart Contracts Solidity, SQL, GraphQL, Protocol Buffers (Protobuf), Prisma
Manifests & Builds package.json, pyproject.toml, setup.py, requirements.txt, Makefile, Cargo.toml, go.mod, pom.xml, build.gradle, Gemfile, composer.json, CMakeLists.txt, pubspec.yaml, Package.swift, mix.exs, Dockerfile, .env.example

CLI Command Reference

# Analyze current directory:
skilly .

# Analyze specific project directory:
skilly /path/to/project

# Serve interactive knowledge graph in browser:
skilly . --serve --port 8080

# Output artifacts to custom directory:
skilly . -o ./docs

# Generate only capability catalog or only knowledge graphs:
skilly . --skills-only
skilly . --graph-only

# Multi-threaded parallel file analysis:
skilly . --workers 8

# Quality gates for CI/CD:
skilly . --fail-on-cycles --fail-on-grade=A

Automated Testing

Skilly contains a comprehensive test suite covering all AST parsers, NetworkX algorithms, incremental caching, and CLI quality gates:

python -m pytest tests/ -v
============================= 20 passed in 1.21s ==============================

Author

Skilly was conceptualized, designed, and engineered from the ground up by Arastu Thakur.

Arastu Thakur Arastu Thakur
Data Scientist

Website: arastuthakur.com.np
GitHub: @arastuthakur
LinkedIn: in/arastuthakur
PyPI Package: skilly-ai
Email: arustuthakur@gmail.com

License

This project is licensed under the MIT License — see the LICENSE file for details.

Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distribution

skilly_ai-1.0.1.tar.gz (61.8 kB view details)

Uploaded Source

Built Distribution

If you're not sure about the file name format, learn more about wheel file names.

skilly_ai-1.0.1-py3-none-any.whl (65.2 kB view details)

Uploaded Python 3

File details

Details for the file skilly_ai-1.0.1.tar.gz.

File metadata

  • Download URL: skilly_ai-1.0.1.tar.gz
  • Upload date:
  • Size: 61.8 kB
  • Tags: Source
  • Uploaded using Trusted Publishing? No
  • Uploaded via: twine/7.0.0 CPython/3.13.14

File hashes

Hashes for skilly_ai-1.0.1.tar.gz
Algorithm Hash digest
SHA256 b1c29cf783f884c41f7791f17e4f9e65d51190fa7b234c255ce3cfcd62a7edbf
MD5 f79d5afbc1387b4c25ff41416c93f261
BLAKE2b-256 1e298e9b0ed9fd94f30ba47358c76aab992f51d7a5aec553855cfe4c0848e182

See more details on using hashes here.

File details

Details for the file skilly_ai-1.0.1-py3-none-any.whl.

File metadata

  • Download URL: skilly_ai-1.0.1-py3-none-any.whl
  • Upload date:
  • Size: 65.2 kB
  • Tags: Python 3
  • Uploaded using Trusted Publishing? No
  • Uploaded via: twine/7.0.0 CPython/3.13.14

File hashes

Hashes for skilly_ai-1.0.1-py3-none-any.whl
Algorithm Hash digest
SHA256 f97795de8da7cad8068ff697017198fc1e4d6b4aed0474998c0556e5b0c9279d
MD5 ddc6117621e59f9522db33300369a2b4
BLAKE2b-256 fb8f5896c9e8df44cdd6d49a61299efa750dcaf43d3721dfdf71029d4dee94de

See more details on using hashes here.

Release history Release notifications | RSS feed

1.0.4

2 files

1.0.3

2 files

1.0.2

2 files

This release

1.0.1 This release

2 files

1.0.0

2 files

Anthropic, PBC Visionary sponsor Bloomberg Visionary sponsor Hudson River Trading Visionary sponsor Meta Visionary sponsor NVIDIA Visionary sponsor Microsoft Sustainability sponsor Depot Continuous Integration AWS Cloud computing and Security Sponsor Datadog Monitoring Fastly CDN Google Download Analytics Sentry Error logging StatusPage Status page