Best Claude skills for superpowers.
163 skills in ahel's catalog match "superpowers". This list ranks 25 of them by GitHub stars, with the ones named for superpowers first. The top three are brainstorming, dispatching-parallel-agents and executing-plans. They come from 7 different publishers. All 25 have GitHub stars on record, and brainstorming has the most, 293,515. 14 of the 25 show weekly installs, and systematic-debugging leads with 8,538 a week. All 25 come from the community rather than from the vendor of the product they work with. 23 of the 25 were read in full by ahel's safety review. All 25 work through ahel today, so one connection brings them to Claude, ChatGPT and Ahel Desktop. All 25 had a commit in the last 90 days. Most sit in Dev tools and Productivity. Each row links to the item's own page, with its install steps and safety review.
25 results
| # | Name | Stars | Installs a week | Publisher | Use in |
|---|---|---|---|---|---|
| 1 | brainstormingAdds a brainstorming skill that makes your agent explore intent and design before writing any code. | 294k | 7k | Communityobra | ClaudeChatGPTAhel Desktop |
| 2 | dispatching-parallel-agentsLets your agent split independent tasks and run several sub-agents on them at the same time. | 294k | 5k | Communityobra | ClaudeChatGPTAhel Desktop |
| 3 | executing-plansLets your agent execute an implementation plan itself using the executing plans skill when no subagent tool is available. | 294k | 6k | Communityobra | ClaudeChatGPTAhel Desktop |
| 4 | finishing-a-development-branchGuides your agent through finishing a development branch skillfully, choosing merge, pull request, or cleanup. | 294k | 5k | Communityobra | ClaudeChatGPTAhel Desktop |
| 5 | receiving-code-reviewGuides your agent in receiving code review skill feedback with verification instead of blind agreement. | 294k | 5k | Communityobra | ClaudeChatGPTAhel Desktop |
| 6 | requesting-code-reviewGives your agent a requesting code review skill to check its finished work meets requirements before merging. | 294k | 6k | Communityobra | ClaudeChatGPTAhel Desktop |
| 7 | subagent-driven-developmentLets your agent apply the subagent driven development skill to execute implementation plans with independent tasks. | 294k | 6k | Communityobra | ClaudeChatGPTAhel Desktop |
| 8 | systematic-debuggingGives your agent a systematic debugging claude skill for digging into bugs and test failures before fixing. | 294k | 9k | Communityobra | ClaudeChatGPTAhel Desktop |
| 9 | test-driven-developmentLets your agent write failing tests first and code second using a test driven development skill. | 294k | 6k | Communityobra | ClaudeChatGPTAhel Desktop |
| 10 | using-git-worktreesLets your agent create an isolated workspace using git worktrees skill before starting feature work. | 294k | 5k | Communityobra | ClaudeChatGPTAhel Desktop |
| 11 | using-superpowersLets your agent use the using superpowers skill to find and apply its skills before responding to any request. | 294k | 5k | Communityobra | ClaudeChatGPTAhel Desktop |
| 12 | verification-before-completionMakes your agent run verification commands and check output before claiming any verification before completion skill. | 294k | 7k | Communityobra | ClaudeChatGPTAhel Desktop |
| 13 | writing-plansLets your agent turn a spec or requirements into a step-by-step plan before any code is written. | 294k | 7k | Communityobra | ClaudeChatGPTAhel Desktop |
| 14 | writing-skillsLets your agent create and edit new skills and check they work before use. | 294k | 5k | Communityobra | ClaudeChatGPTAhel Desktop |
| 15 | chinese-code-reviewLets your agent phrase code review comments in Chinese using severity labels and common review templates. | 8k | none | Communityjnmetacode | ClaudeChatGPTAhel Desktop |
| 16 | chinese-commit-conventionsLets your agent write commit messages and changelogs following Chinese-language conventions. | 8k | none | Communityjnmetacode | ClaudeChatGPTAhel Desktop |
| 17 | chinese-documentationLets your agent follow Chinese documentation style rules for spacing, punctuation, and copywriting. | 8k | none | Communityjnmetacode | ClaudeChatGPTAhel Desktop |
| 18 | chinese-git-workflowGives your agent reference info on configuring Git access and CI for Chinese platforms like Gitee and JiHu GitLab. | 8k | none | Communityjnmetacode | ClaudeChatGPTAhel Desktop |
| 19 | mcp-builderGuides your agent through designing, building, and testing server tools that connect AI assistants to external services. | 8k | none | Communityjnmetacode | ClaudeChatGPTAhel Desktop |
| 20 | workflow-runnerLets your agent run multi-role YAML workflows using the current AI session itself as the execution engine. | 8k | none | Communityjnmetacode | ClaudeChatGPTAhel Desktop |
| 21 | Getting Started with Research SuperpowersGuides your agent through finding, screening, and extracting data from scientific papers step by step. | 4k | none | Communitybrycewang-stanford | ClaudeChatGPTAhel Desktop |
| 22 | using-startup-superpowersUse at the start of any conversation about a startup idea, product validation, founder strategy, or work inside a `startup/` workspace. Establishes file conventions, voice-input handling, subagent dispatch rules, and how to update each artifact safely. Activate before invoking any other startup-supe | 4k | none | Communitydavepoon | ClaudeChatGPTAhel Desktop |
| 23 | using-bgs-modding-superpowersLets your agent load Bethesda modding rules and a task skill list at the start of any modding conversation. | 1k | none | Communityhashgraph-online | ClaudeChatGPTAhel Desktop |
| 24 | diagnosing-superpowersLets your agent investigate why a superpowers session went wrong and build a bug report. | 973 | none | Communityguanyang | ClaudeChatGPTAhel Desktop |
| 25 | designing-the-analysisHelps your agent turn an approved research question into a concrete analysis plan before any data work starts. | 344 | none | Communityk-dense-ai | ClaudeChatGPTAhel Desktop |