Tooling & MCP
nanocoai/nanoclaw
flagship skill dial-cli
core
Skill Radar
Updated 2 Nov 2026, 16:51 UTC · 324 setups tracked · 16 capability groups
Which agent skills are actually being adopted, ranked by momentum and adoption rather than raw stars. A setup is one repository or a family of ports of it. Start with the proven ones, then browse by the job you need done.
Proven setupstop 6 of 24
Tooling & MCP
paperclipai/paperclip
flagship skill create-external-bot-integration
core
Documents & Files
anthropics/financial-services
flagship skill ic-memo
adjacent
Security
koala73/worldmonitor
flagship skill monitor-supply-chain-stress
core
Testing
garrytan/gstack
flagship skill ios-qa
core
Writing
coreyhaines31/marketingskills
flagship skill copywriting
core
All 24 setups, side by side
| Setup | Main capability | Groups | Skills | Forks | Stars |
|---|---|---|---|---|---|
| nanocoai/nanoclawcore | Tooling & MCP | 4 | 64 | 40.8% | 31.1k |
| paperclipai/paperclipcore | Tooling & MCP | 8 | 76 | 16.6% | 109.9k |
| anthropics/financial-servicesadjacent | Documents & Files | 2 | 64 | 13.9% | 41.0k |
| koala73/worldmonitorcore | Security | 3 | 37 | 15.4% | 89.9k |
| garrytan/gstackcore | Testing | 5 | 68 | 14.8% | 138.6k |
| coreyhaines31/marketingskillscore | Writing | 4 | 53 | 14.7% | 57.8k |
| cheaper-inference/OmniRoutecore | Tooling & MCP | 1 | 52 | 14.9% | 83.6k |
| HKUDS/Vibe-Tradingadjacent | Research | 2 | 91 | 16.2% | 37.4k |
| openclaw/clawhubcore | Infra & Deploy | 6 | 67 | 15.7% | 9.5k |
| rohitg00/ai-engineering-from-scratchcore | Code Review | 3 | 23 | 17.7% | 76.2k |
| aden-hive/hivecore | Debugging | 3 | 24 | 50.7% | 11.1k |
| anthropics/skillsadjacent | Documents & Files | 3 | 20 | 11.8% | 186.3k |
| affaan-m/ECCcore | Frontend & Design | 14 | 317 | 14.8% | 294.8k |
| github/awesome-copilotcore | Tooling & MCP | 17 | 400 | 12.9% | 40.7k |
| garrytan/gbrainadjacent | Research | 6 | 80 | 14.9% | 31.4k |
| wshobson/agentscore | Frontend & Design | 15 | 185 | 10.6% | 42.1k |
| AgriciDaniel/claude-seocore | Data & SQL | 1 | 33 | 13.2% | 25.5k |
| pingdotgg/t3codecore | Debugging | 2 | 15 | 23.6% | 37.8k |
| Yeachan-Heo/oh-my-claudecodeadjacent | Research | 3 | 49 | 8.9% | 40.5k |
| vudovn/ag-kitcore | Frontend & Design | 6 | 58 | 18.4% | 8.2k |
| Donchitos/Claude-Code-Game-Studioscore | Testing | 9 | 74 | 13.6% | 32.2k |
| CherryHQ/cherry-studiocore | Code Review | 2 | 40 | 9.7% | 53.2k |
| code-yeongyu/oh-my-openagentcore | Testing | 6 | 47 | 8.2% | 70.6k |
| AgriciDaniel/claude-adsadjacent | Research | 1 | 34 | 14.5% | 10.7k |
Browse by capability16 groups
Tooling & MCPgoogle/agents-cliagents-cli-claudeTestingWaishnav/devspacereplay-qaDebuggingcobusgreyling/loop-engineeringci-triageInfra & Deploymicrosoft/azure-skillsazure-deployEvals & BenchmarksOpenRaiser/NanoResearchlm-evaluation-harnessRefactor & ArchitectureJuliusBrussee/cavemanmigrationCode ReviewjnMetaCode/superpowers-zhchinese-code-reviewMemory & Contextmichael-denyer/pstack-claudeprinciple-guard-the-context-windowOrchestrationamElnagdy/delegate-skillsagy-delegateGit & VCSMoonshotAI/kimi-codepre-changelogData & SQLKaelio/ktxdbt_ingestSecurityopenai/codex-securitythreat-modelCodebase Navigationwithcoral/coralcodebase-designDocuments & Fileseigent-ai/eigentdocxDev Environmentmindfold-ai/Trellisbootstrap-skillResearchImbad0202/academic-research-skillsacademic-paper
Rising this week4
- DenisSergeevitch/ex-ussr-kinship-researchfirst seen 2026-11-01
- Fei2411/DS-Botfirst seen 2026-11-01
- ccch1mneyyy/dsh-Gofirst seen 2026-11-01
- daveebbelaar/bamafirst seen 2026-11-01
Practitioners ranked by ports (12)
A port is someone else re-implementing the same setup for another harness, so it is the strongest adoption signal we can observe.
| Setup | Author | Ports | Stars |
|---|---|---|---|
| kepano/obsidian-skills +2 ports | kepano | 2 | 72.1k |
| rtk-ai/rtk +1 port | rtk-ai | 1 | 117.0k |
| vercel-labs/native +2 ports | vercel-labs | 2 | 18.0k |
| michael-denyer/pstack-claude +1 port | michael-denyer | 1 | 7.0k |
| cheaper-inference/OmniRoute | cheaper-inference | 0 | 83.6k |
| mattpocock/skills | mattpocock | 0 | 309.2k |
| coreyhaines31/marketingskills | coreyhaines31 | 0 | 57.8k |
| wshobson/agents | wshobson | 0 | 42.1k |
| deepseek-ai/deepseek-harness | deepseek-ai | 0 | 268.3k |
| K-Dense-AI/scientific-agent-skills | K-Dense-AI | 0 | 55.7k |
| JuliusBrussee/caveman | JuliusBrussee | 0 | 116.7k |
| anthropics/financial-services | anthropics | 0 | 41.0k |
How this is ranked
- Discovery is a union of GitHub repo searches (new-and-fast repos, topic-tagged repos, and name-family ports), never a single query.
- One skill is counted once even when a repo vendors it for many harnesses; the copy count is portability, not popularity.
- Stars are shown but never ranked. Adoption is fork ratio, because installing a skill means copying it.
- Ranking families lets a ported setup be measured as one concept instead of many small repos.
- A capability group is shown only when at least two skills in the setup name that capability, and the most focused setup wins the group: two matches out of five skills beats four out of four hundred.
- Core means the capability serves agent and platform engineering; adjacent means useful but broader. Off-lens setups are not listed at all.
- 419 setups were excluded for being outside the platform/agent-engineering lens.