agent-skills vs agentic-engineering
Workflows for AI coding agents compared
agent-skills and agentic-engineering both provide Markdown workflows to guide AI coding agents. agent-skills spans more development lifecycle tasks and supports several named tools, while agentic-engineering focuses on documentation, persistent project context, and independent verification across sessions.
agent-skills: Production-Grade Workflows for AI Coding Agents
A portable pack of Markdown engineering skills that helps AI coding agents follow senior-engineer workflows from scoping through shipping. It supports a range of agents, including Claude Code, Cursor, Codex, Copilot, and Gemini CLI.
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-skills | agentic-engineering | |
|---|---|---|
| Language | JavaScript | — |
| License | MIT | MIT |
| Stars | 109k | 5 |
| Forks | 11.3k | 0 |
| Last analyzed | Oct 20, 2026 | Oct 20, 2026 |
Key differences
- agent-skills offers 25 skills and nine Claude Code lifecycle commands; agentic-engineering organizes ten skills around documentation, planning, tracking, and verification.
- agent-skills emphasizes reusable engineering practices across specification, implementation, testing, security, review, and shipping; agentic-engineering centers on repository documents such as specifications, decision records, and status tracking.
- agent-skills lists setup for Claude Code, Cursor, Codex, Copilot, Gemini CLI, and Antigravity CLI; agentic-engineering is described for Claude Code, Codex, and compatible agents, with setup potentially needed for other tools.
- agentic-engineering separates implementation from fresh-context verification and its loop driver can prepare draft PRs, but a human reviews and merges; agent-skills emphasizes explicit evidence and exit criteria across workflows.
- agent-skills is actively maintained and reports 109k stars and 11.3k forks; agentic-engineering reports 5 stars and 0 forks and has had no commits in the last 90 days.
- Both projects use the MIT license; agent-skills lists JavaScript as its language, while agentic-engineering describes its skills as Markdown playbooks.
Choose agent-skills if you…
- need workflows spanning requirements, implementation, testing, security, review, and shipping.
- want to share engineering practices across several named coding-agent tools.
- value a larger skill pack with specialist personas and review checklists.
Choose agentic-engineering if you…
- need project knowledge, decisions, and task status persisted in repository documents across sessions.
- want implementation checked by a separate, fresh-context verifier before a PR is ready.
- need a workflow that can create task branches and draft PRs while retaining human review and merge control.
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.