agent-skills vs superpowers
Structured workflows for AI coding agents compared
agent-skills and superpowers both provide reusable skills and workflows for guiding coding agents through software development. agent-skills emphasizes specification-to-PR processes and an optional autonomous sprint harness, while superpowers provides a more prescriptive workflow with test-driven development and branch completion.

agent-skills: Structure AI-Assisted Development from Specs to PRs
A collection of reusable skills for AI coding agents, covering specification writing, implementation, review, testing, and autonomous sprint workflows. Useful for teams that want repeatable processes and clearer quality gates across agent sessions.

superpowers: Add a Software Development Methodology to Coding Agents
Superpowers is a composable skills framework and software development methodology for coding agents. It guides work from requirements and planning through test-driven implementation and review, with integrations for multiple coding-agent platforms.
| agent-skills | superpowers | |
|---|---|---|
| Language | Shell | Shell |
| License | MIT | MIT |
| Stars | 2 | 308k |
| Forks | 0 | 27.4k |
| Last analyzed | Oct 3, 2026 | Oct 29, 2026 |
Key differences
- agent-skills focuses on creating and checking requirements, designs, task lists, and acceptance test plans; superpowers emphasizes brainstorming, planning, test-driven implementation, and finishing branches.
- agent-skills includes workflows for delegating tasks, independent reviews, issue-to-PR orchestration, and an optional autonomous sprint harness; superpowers offers subagent or inline execution and worktree support.
- agent-skills can convert Claude Code MCP settings to Codex CLI configuration and suggest skills from project files; superpowers includes session diagnosis that can prepare a scrubbed report.
- superpowers describes its workflow as prescriptive, with mandatory skills and test-driven development; agent-skills supports manual approval or automated execution within human-defined boundaries.
- Both projects use Shell and have MIT licenses. superpowers has 295k stars and 26.3k forks, while agent-skills has 2 stars and 0 forks.
- Both require a compatible coding-agent setup; superpowers requires separate installation for each harness, while agent-skills skills support agents that implement the SKILL.md format.
Choose agent-skills if you…
- need specification writing, validation, and task-based implementation workflows.
- want to orchestrate work from specifications or issues through pull-request preparation.
- want to trial an optional autonomous sprint workflow with explicit project boundaries.
Choose superpowers if you…
- want a guided process from brainstorming and planning through test-driven implementation and branch completion.
- prefer task execution using subagents or inline work with reviews.
- need skills for systematic debugging, worktrees, or diagnosing an agent session.
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.