📈 GitHub Trending - Daily

Last updated: 2026-02-06 01:10 UTC

Repository	⭐ Stars	Language	Description

| bytedance/UI-TARS-desktop | 26704 | TypeScript | The Open-Source Multimodal AI Agent Stack: Connecting Cutting-Edge AI Models and Agent Infra |

| openai/skills | 4277 | Python | Skills Catalog for Codex |

| thedotmack/claude-mem | 23714 | TypeScript | A Claude Code plugin that automatically captures everything Claude does during your coding sessions, compresses it with AI (using Claude's agent-sdk), and injects relevant context back into future sessions. |

| j178/prek | 5514 | Rust | ⚡ Better pre-commit, re-engineered in Rust |

| topoteretes/cognee | 11837 | Python | Memory for AI Agents in 6 lines of code |

| obra/superpowers | 45361 | Shell | An agentic skills framework & software development methodology that works. |

| aquasecurity/trivy | 31367 | Go | Find vulnerabilities, misconfigurations, secrets, SBOM in containers, Kubernetes, code repositories, clouds and more |

| fish-shell/fish-shell | 32385 | Rust | The user-friendly command line shell. |

| nvm-sh/nvm | 91373 | Shell | Node Version Manager - POSIX-compliant bash script to manage multiple active node.js versions |

| linshenkx/prompt-optimizer | 19350 | TypeScript | 一款提示词优化器，助力于编写高质量的提示词 |

TrendSpire

TrendSpire gathers trending repositories from GitHub and stores them in TRENDING.md. GitHub Actions keep the digest fresh and leverage OpenAI Codex to continuously improve the codebase.

Features

• Automated scraping of GitHub's trending page with configurable language, time range and result limit.
• Daily workflow to regenerate TRENDING.md and update this README.
• Scheduled Codex runs that suggest small refactors and new tests via pull requests.
• Token and cost tracking for all Codex requests.
• Persistent memory stored under trendspire_memory/ enables the AI to iteratively refine its suggestions across runs and open automated pull requests with context.

What TrendSpire Does Today

• python -m src.fetch_trending — scrape GitHub Trending
• python -m src.render_digest — render TRENDING.md & inject into README.md

AI Agents (coming soon)

See AGENTS.md and ai_loop/README.md for details on the self-improvement loop.

Getting Started

1. Install dependencies

   python3 -m venv venv
   source venv/bin/activate
   pip install -r requirements.txt

2. Run the setup wizard

   python scripts/setup_wizard.py

   This interactive script stores your preferred trending options and OpenAI API key. You can rerun it at any time to change the configuration.

3. Run the trending scraper

   python -m src.render_digest

   The latest results will appear in TRENDING.md and the README.

GitHub Actions

Update Digest

The workflow update_digest.yml runs every day at 08:00 UTC. It installs the dependencies, executes python -m src.render_digest, and commits any changes to TRENDING.md and README.md.

Codex Automation

Another workflow ai_loop.yml drives the Codex automation using ai_loop/autoloop.py. It supports two modes:

• Daily – diff-based improvements using gpt-3.5-turbo.
• Weekly – a full repository review with gpt-4o.

Each run applies the returned diff, executes the test suite and, when successful, creates a branch and pull request. Summaries, cost logs and the raw diff are saved under codex_logs/ and uploaded as workflow artifacts. The workflow also caches the trendspire_memory/ directory so the AI can refine its suggestions over time.

To run the Codex automation locally you can execute:

python -m ai_loop.autoloop

API usage reports

The file logs/api_usage.* records token counts and cost. Set API_LOG_FORMAT to csv, json or txt to control the format. Run python scripts/summarize_usage.py for a quick summary grouped by model.

Running tests

After installing the requirements you can run the entire test suite with

pytest

Additional tips for contributors are available in docs/DEVELOPER.md.

🗃 Archived & Legacy Code

This repo includes experimental or deprecated files that are not part of the active AI loop. These are stored in:

• legacy/ – old logic and patch tools
• archive/ – past metrics and planning reports
• later/ – utilities planned for future releases (Phase 5+)