5 Open-Source GitHub Repos for Agentic Workflows

Agentic Tooling and Workflow Automation

The author highlights five open-source repositories that provide functional capabilities for AI agents, moving beyond simple chat interfaces toward task-oriented execution. The selection focuses on tools that can be productized or used to build internal leverage.

Repository Breakdown

• no-ai-slop: An editor skill designed to strip formulaic AI patterns from text while preserving the user's original voice. It is intended to be used as a post-processing step after a human has drafted the core ideas.
• crm: An agent-first CRM that functions as a workspace rather than a static database. It automates lead enrichment, follow-up scheduling, and relationship graph maintenance. Setup requires bun and docker, with local hosting on localhost:3000 and localhost:3001 for the API.
• video-use: A framework allowing coding agents to perform video editing tasks such as removing filler words, color grading, and adding overlays. The author recommends starting with a single repeatable format, such as converting a founder's Loom video into a 60-second launch clip.
• SkillSpector: A security scanner from NVIDIA that inspects AI agent skills for prompt injection, data exfiltration, and supply chain risks. It supports a static scan mode via the --no-llm flag to avoid sending sensitive file contents to external providers.
• phone-harness: A tool that enables agents to control physical mobile devices. It utilizes Mac iPhone mirroring for iOS and ADB for Android, allowing agents to tap, type, and scroll through apps that lack public APIs for QA or automation purposes.

Implementation Strategy

The author advocates for a three-step workflow to evaluate these tools: install the repository, build one small, high-value workflow, and then decide whether to productize the result or keep it as internal leverage. This approach emphasizes learning through installation and iterative testing rather than attempting full-scale automation immediately.