GitHub Trending weekly on October 25, 2026

A coding-agent skill that runs multi-phase security audits and writes verified, schema-checked findings.

An orchestration desktop for running many coding agents in parallel git worktrees, with phone monitoring and remote execution.

Builds a local, queryable knowledge graph of a codebase so coding agents can navigate it instead of grepping.

An MIT-licensed terminal coding agent for developers who want full control over model choice and local file access.

Turns HTML, CSS and seekable animations into deterministic MP4s for developers and agents producing video programmatically.

Node-graph engine for running open diffusion image, video and audio models locally with fine control over every step.

Shared long-term memory for coding agents so you can switch harnesses or machines mid-task without re-explaining context.

A harness that wraps Claude Code and Codex with swarms, persistent memory, and cross-machine federation for multi-agent work.

A curated plugin marketplace for Claude Code, giving developers a vetted place to find and install extensions instead of hunting GitHub.

Generated command-line harnesses that let AI agents drive desktop apps that only have a GUI.

A CLAUDE.md rule set for coding agents that pushes them to ask before assuming, keep diffs small, and work toward verifiable goals.

Role-specific Claude Cowork plugins that give non-engineers prebuilt skills, commands and tool connectors for their job function.

Control Claude Code, Codex and other coding agents from a local server, web app, desktop app or phone.

A plugin marketplace and spec for giving Cursor's coding agent reusable skills and third-party service integrations.

Serves and batches LLM inference at high throughput for teams running their own models.

Official Android skill files that teach coding agents how to handle Android tasks LLMs usually get wrong.