agentic-engineering vs agent-feed
Repository workflows for AI coding agents compared
agentic-engineering and agent-feed both organize AI-assisted coding around repository instructions, verification, and work that can span sessions. agentic-engineering is a documentation-first skill suite, while agent-feed is a Python CLI that installs and checks a shared workflow with adapters for multiple coding tools.
agentic-engineering: Give Coding Agents a Documentation-First Workflow
A documentation-first harness of ten chain skills for Claude Code, Codex, and compatible coding agents. It provides persistent project memory, implementation contracts, decision records, task planning, and independent verification for long-running development.
agent-feed: Govern AI Coding Workflows in Your Repository
Agent Feed installs a repository-owned workflow for AI coding tools, with shared instructions, task and review gates, checks, and tool adapters. It suits teams that want consistent, verifiable AI-assisted changes across Codex, Claude Code, and Cursor.
| agentic-engineering | agent-feed | |
|---|---|---|
| Language | — | Python |
| License | MIT | MIT |
| Stars | 5 | 4 |
| Forks | 0 | 0 |
| Last analyzed | Oct 20, 2026 | Oct 12, 2026 |
Key differences
- agentic-engineering provides skills and workflow guidance; agent-feed provides a CLI to install, configure, and validate workflow assets.
- agentic-engineering emphasizes project specifications, decision records, task checkpoints, and independent verification; agent-feed emphasizes task boundaries, decision and review gates, checks, and tool adapters.
- agentic-engineering describes a loop driver that advances planned tasks through build, verification, and checkpoints; agent-feed includes commands for status, upgrades, adapter sync, and configuration.
- agentic-engineering describes a Git workflow with one branch and draft PR per task, leaving merges to humans; agent-feed describes review gates and human confirmation for some decisions.
- agentic-engineering is documented for Claude Code and manual use with Codex and other tools; agent-feed provides adapters for Codex, Claude Code, and Cursor.
- Both use the MIT license. agentic-engineering has 5 stars and agent-feed has 4; agent-feed also notes a single contributor and a quiet recent measured period.
Choose agentic-engineering if you…
- want a documentation-first process with PRDs, implementation specifications, and append-only decision records.
- need persistent task checkpoints and independent, evidence-based verification across sessions.
- want a loop driver for planned work while keeping merges reserved for humans.
Choose agent-feed if you…
- use Codex, Claude Code, and Cursor and want a shared repository workflow across them.
- need a Python CLI to install and validate workflow instructions and supporting checks.
- want to manage reusable and imported skills alongside project-specific constraints and knowledge.
Compare the numbers side by side, or add more projects
This comparison is generated with AI from the OSRepos analyses of both projects. Always check each project's repository and documentation before choosing.