Skillget

Codentis

agent
Security Audit
Warn
Health Warn
  • No license — Repository has no license file
  • Description — Repository has a description
  • Active repo — Last push 0 days ago
  • Community trust — 10 GitHub stars
Code Pass
  • Code scan — Scanned 12 files during light audit, no dangerous patterns found
Permissions Pass
  • Permissions — No dangerous permissions requested
Purpose
This is an AI-powered terminal agent that analyzes your codebase and assists with development tasks. It interacts directly with your local files to provide code reviews, debugging, and automated workflows.

Security Assessment
The tool requires access to local file systems to read and analyze code. Because it is an AI agent, it inherently makes network requests to external LLM providers (like OpenAI or Anthropic) to function, meaning your codebase queries will be sent externally unless you configure a local, offline LLM endpoint. The automated code analysis suggests it likely executes or evaluates commands within a terminal environment. The automated scan found no hardcoded secrets or dangerous code patterns, and the tool requires no dangerous system permissions. Overall risk is rated as Medium due to the standard privacy implications of sending local code context to third-party AI APIs.

Quality Assessment
The project appears to be actively maintained, with repository pushes occurring as recently as today. The developer notes it is licensed under the MIT License, though the automated scanner flagged a missing license file in the repository directory, meaning the legal terms are currently undocumented in the standard format. Community trust is very early-stage, represented by only 10 GitHub stars. The codebase passed a basic security scan cleanly across its 12 files.

Verdict
Use with caution: The code itself appears safe and actively maintained, but early adoption status and the standard privacy implications of sending your codebase to third-party AI servers warrant careful configuration.
SUMMARY

Codentis is an AI-powered terminal tool that understands your codebase, analyzes repositories, and helps you explore and work with code faster.

README.md

Codentis

Codentis is a high-performance CLI AI agent designed for terminal-based developers. It provides intelligent code assistance, file operations, and automated workflows through a modular agentic loop.

Distinguishing Factors

  • Sub-Agent Delegation: Automatically offloads complex analysis, debugging, and refactoring to specialized sub-agents with curated goal prompts.
  • Structured Tool Registry: Features collapsible outputs with persistent IDs, allowing developers to expand only the data they need to see.
  • Workspace Security: Implements a proactive trust system requiring explicit user approval before performing operations in new directories.
  • Provider Agnostic: Seamlessly switches between OpenAI, Anthropic, OpenRouter, and local OpenAI-compatible endpoints.
  • Native Packaging: Ships as standalone installers (EXE, PKG, DEB) for professional deployment beyond standard pip environments.

Use Cases

  • Codebase Onboarding: Build a comprehensive knowledge graph and dependency map of unfamiliar projects in a single pass.
  • Automated Code Audits: Perform deep security and performance reviews using specialized reviewer agents before committing changes.
  • Legacy Migration: systematically translate logic between frameworks or language versions with correctness verification.
  • Local Development: Connect to local LLM servers for private, offline-first development workflows.

Installation

From Release (Recommended)

Download the latest installer for your platform from the Releases page.

Development

git clone https://github.com/sujal-GITHUB/Codentis.git
cd Codentis
pip install -e .
codentis

Usage

Start an interactive session:

codentis

Run a single prompt:

codentis chat "Analyze the current project structure"

Common TUI commands:

  • /e <id>: Expand specific tool output
  • /list: Show all tool calls in session
  • /exit: Quit the session

Configuration

Manage settings with the config command:

codentis config --show
codentis config --reset

The application stores settings in ~/.codentis/config.json. Project-specific overrides can be defined in .agent/codentis.toml.

Architecture

Codentis follows a modular design:

  • agent/: Core agent loop and session management.
  • client/: LLM provider integrations (OpenAI, Anthropic, OpenRouter).
  • tools/: Extensible tool registry and implementations.
  • ui/: Terminal rendering and user interaction.

Development

Run tests:

pytest

Build native installers:

python scripts/build_exe.py 64  # Windows
./scripts/build_macos.sh         # macOS
./scripts/build_linux.sh         # Linux

License

This project is licensed under the MIT License.

Acknowledgments

Inspired by modern AI-powered developer tools like Claude Code.

Reviews (0)

No results found