Find the open-source solution that fits.

Project overview

Describe a time range, scene, or output goal in Claude Code to cut footage, remove silence, turn long videos into Shorts, or generate and style subtitles. Changes are previewed through a dry run and backed by project backups and rollback guidance. The skill depends on the unofficial capcut-cli and CapCut or JianYing Desktop, so CapCut Web is unsupported and application schema changes may break compatibility.

Project overview

This local-first skill helps an agent lead a GA4 and Google Tag Manager project from site analysis through review and delivery. It groups pages by business purpose, drafts or audits an event schema, compares it with a live GTM setup, and can continue into generation, synchronization, and preview verification only within the requested scope. It supports generic sites and a separate Shopify path. Browser-backed steps use Playwright, GTM synchronization uses Google OAuth, and richer anonymous diagnostics remain opt-in beyond a limited startup signal.

Project overview

rsi-loop selects the highest-priority unchecked medium task from BACKLOG.md, implements the smallest correct change, runs the repository’s tests, typecheck, and build, then commits it. When the backlog is empty it can regenerate more work, and .rsi-loop.json controls the commit target, perpetual mode, and blast-radius limit. Because it has no question or checkpoint step, use it only where the task boundaries and checks are already trustworthy.

Project overview

loop-creator helps decide whether recurring work deserves a bounded agent loop, adapts one of 50 patterns, or designs a new loop through a plain-language interview. It can diagnose missing brakes, reward hacking, self-approval, and stale state, with a separate checking instance and explicit stop or human-question conditions.

Project overview

Image Prompt Generator turns a visual brief into three concepts, delegates style and booru reference research, then writes final prompts and runs parallel image generation and review. It is built around Codex image generation, with Python and gallery-dl for reference collection, while also adapting prompts for Midjourney, Stable Diffusion, Flux, and other tools.

Project overview

atomizekit turns a topic into a blog post, native drafts for seven platforms, branded image cards, and voiced short-form video, then gathers the files in a local review dashboard. It configures itself from the host project's brand assets and can track gaps in AI-answer citations. Everything remains draft-only for human posting, and video output requires tools such as FFmpeg and Chrome.

Project overview

This agent skill helps consultants and business teams turn a company name into an AI opportunity map. It guides a web-search-capable agent through company and industry research, pain-point analysis, scenario discovery, prioritization, and implementation planning, with quick, standard, and deep report modes. The repository includes industry examples, analysis frameworks, and report templates. Research quality depends on the connected agent and its search and reasoning capabilities, and the supporting framework script requires Python 3.8 or later.

Project overview

Untether bridges coding agents running on your own computer or server to a Telegram bot. From a phone, you can send text or voice tasks, stream progress, switch projects and sessions, transfer files, and respond when an agent needs input. It supports Claude Code, Codex, OpenCode, Pi, Gemini CLI, and Amp, but approval controls, plan mode, and cost reporting vary by engine.

Project overview

claude-loop converts “make it better” into a project-specific scorecard and a repeatable cycle: score real signals, triage the weakest link, fix one slice, verify independently, and score again. Its init, score, run, and status modes keep the rubric and checkpoint state in .loop/. The loop stops after the target score is reached twice, with 90 as the default, instead of running indefinitely.

Project overview

IRIS gives Claude Code a disciplined network-operations workflow for discovering device state, checking it against a source of truth, preparing a change, verifying the result, and documenting the ticket. Engineers retain the decision, while connected MCP tools, network documents, and discovery scripts supply context. Read-only automation can assist investigation; state-changing actions require explicit approval.

Project overview

agent-terminal lets an AI agent launch interactive CLI applications, capture their screen as ASCII text, and send keyboard input without a display. It can be used as an npm library or exposed through an MCP server for CLI testing, monitoring, and build automation. Built-in safeguards include command allowlists, workspace path validation, environment sanitization, input checks, and session limits.

Project overview

Wreckit takes a text backlog and moves each item through codebase research, planning, story-by-story implementation, and pull request creation for you to review. State and artifacts stay in Git-trackable JSON and Markdown files, making runs inspectable and resumable. It requires an AI agent or API access and can consume substantial tokens; an optional Sprite-based sandbox isolates riskier execution.

Project overview

open-campsite-watcher polls recreation.gov for the campground and nights you choose, compares each response with saved state, and sends a Telegram alert when a new cancellation appears. The alert includes a one-tap Add to Cart link, while expiry dates, repeat limits, and trips crossing months are handled for you. It runs as a zero-dependency Python watcher on macOS or Linux; booking remains a manual step.

Project overview

Agent Sandbox Skill gives Claude Code, Codex, and Gemini CLI managed E2B environments for engineering work that should not touch the local filesystem or production systems. Agents can install dependencies, run untrusted code, scaffold and build full-stack applications, expose a hosted port, and perform Playwright-based visual checks. Commands cover small sandbox tasks as well as plan, build, host, and test workflows, with lifecycle tools that preserve a sandbox across turns. An E2B API key is required, and compute and hosted previews depend on the E2B service.

Project overview

Lumate takes a Luma event link, reads the form, fills profile fields, and drafts answers from the user’s real GitHub projects before submitting the registration. The first event helps establish reusable answers, so later registrations can run with fewer questions. It keeps the data local, but users should still verify the drafted responses and final submission because the skill acts on an external event form.

Project overview

This kit converts text-layer PDFs from Traditional to Simplified Chinese while trying to retain page size, images, selectable text, and overall layout. An image-only cover can be replaced in a separate step, and the workflow is also packaged as a Codex skill. The included example PDFs and cover are not automatically relicensed under the code’s MIT license.

Project overview

codex-image-gen drives an interactive Codex pane through tmux, requires an explicit absolute save path, waits for the file, and verifies the image and target aspect ratio with ffprobe or PIL. If the expected file is missing, its helper searches the Codex generated-images cache and can recover base64 data from the newest session rollout JSONL. It is a reliability wrapper around an interactive workflow, not a standalone image model.

Project overview

codex-tui-imagegen drives Codex’s interactive terminal UI inside a tmux pane, waits for the generated PNG to appear in the local image directory, and copies it to the path you specify. It is designed for cases where headless execution cannot access image generation or fails to expose a usable file path. The workflow uses an existing Codex/ChatGPT plan instead of a separate image API, but needs an interactive local session.