Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Search, install, and run multi-skill automations from clawflows.com. Combine multiple skills into powerful workflows with logic, conditions, and data flow between steps.
.claude/skills/sundial-org-clawflows/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-01 | ✗→✓ | ▲ Improved | -22% | 0% |
| case-02 | ✗→✓ | ▲ Improved | 17% | 0% |
| case-03 | ✗→✓ | ▲ Improved | -5% | 0% |
| case-04 | ✗→✓ | ▲ Improved | -25% | 0% |
| case-05 | ✗→✓ | ▲ Improved | -30% | 0% |
Discover and run multi-skill automations that combine capabilities like database, charts, social search, and more.
bashnpm i -g clawflows
bashclawflows search "youtube competitor" clawflows search "morning brief" clawflows search --capability chart-generation
Before installing, see what capabilities the automation needs:
bashclawflows check youtube-competitor-tracker
Shows required capabilities and whether you have skills that provide them.
bashclawflows install youtube-competitor-tracker
Downloads to ./automations/youtube-competitor-tracker.yaml
bashclawflows list
bashclawflows run youtube-competitor-tracker clawflows run youtube-competitor-tracker --dry-run
The --dry-run flag shows what would happen without executing.
bashclawflows enable youtube-competitor-tracker # Shows cron setup instructions clawflows disable youtube-competitor-tracker
bashclawflows logs youtube-competitor-tracker clawflows logs youtube-competitor-tracker --last 10
bashclawflows publish ./my-automation.yaml
Prints instructions for submitting to the registry via PR.
Automations use capabilities (abstract) not skills (concrete):
yamlsteps: - capability: youtube-data # Not a specific skill method: getRecentVideos args: channels: ["@MrBeast"] capture: videos - capability: database method: upsert args: table: videos data: "${videos}"
This means automations are portable — they work on any Clawbot that has skills providing the required capabilities.
| Capability | What It Does | Example Skills | |------------|--------------|----------------| | youtube-data | Fetch video/channel stats | youtube-api | | database | Store and query data | sqlite-skill | | chart-generation | Create chart images | chart-image | | social-search | Search X/Twitter | search-x | | prediction-markets | Query odds | polymarket | | weather | Get forecasts | weather | | calendar | Read/write events | caldav-calendar | | email | Send/receive email | agentmail | | tts | Text to speech | elevenlabs-tts |
To make your skill work with ClawFlows automations, add a CAPABILITY.md file:
markdown# my-capability Capability Provides: my-capability Skill: my-skill ## Methods ### myMethod **Input:** - param1: description - param2: description **How to fulfill:** \`\`\`bash ./scripts/my-script.sh --param1 "${param1}" \`\`\` **Output:** Description of output format
And declare it in your SKILL.md frontmatter:
yaml--- name: my-skill provides: - capability: my-capability methods: [myMethod] ---
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | fail→pass | 9,861 | 10,758 | +9% | 1 | 1 | 0% | 1,479 | 1,149 | -22% | 0 | 0 | — |
case-02 | fail→pass | 6,134 | 7,633 | +24% | 1 | 1 | 0% | 970 | 1,134 | +17% | 0 | 0 | — |
case-03 | fail→pass | 11,360 | 5,147 | -55% | 1 | 1 | 0% | 1,780 | 1,699 | -5% | 0 | 0 | — |
case-04 | fail→pass | 23,111 | 1,434 | -94% | 1 | 1 | 0% | 1,343 | 1,008 | -25% | 0 | 0 | — |
case-05 | fail→pass | 11,212 | 2,116 | -81% | 1 | 1 | 0% | 1,562 | 1,100 | -30% | 0 | 0 | — |
case-06 | fail→pass | 16,705 | 2,456 | -85% | 1 | 1 | 0% | 2,597 | 1,139 | -56% | 0 | 0 | — |
case-07 | fail→pass | 11,082 | 2,214 | -80% | 1 | 1 | 0% | 1,652 | 1,094 | -34% | 0 | 0 | — |
case-08 | fail→pass | 8,320 | 1,465 | -82% | 1 | 1 | 0% | 1,364 | 961 | -30% | 0 | 0 | — |
case-09 | fail→pass | 12,903 | 1,327 | -90% | 1 | 1 | 0% | 1,316 | 975 | -26% | 0 | 0 | — |
case-10 | fail→pass | 3,735 | 1,532 | -59% | 1 | 1 | 0% | 509 | 976 | +92% | 0 | 0 | — |
case-11 | fail→pass | 24,409 | 1,703 | -93% | 1 | 1 | 0% | 3,638 | 1,069 | -71% | 0 | 0 | — |
case-12 | pass→pass | 18,591 | 2,192 | -88% | 1 | 1 | 0% | 2,688 | 1,141 | -58% | 0 | 0 | — |
case-13 | fail→pass | 11,380 | 1,627 | -86% | 1 | 1 | 0% | 1,646 | 1,052 | -36% | 0 | 0 | — |
case-14 | pass→pass | 8,293 | 2,180 | -74% | 1 | 1 | 0% | 1,189 | 1,101 | -7% | 0 | 0 | — |
case-15 | fail→pass | 13,935 | 1,520 | -89% | 1 | 1 | 0% | 1,909 | 991 | -48% | 0 | 0 | — |
case-16 | pass→pass | 7,098 | 2,197 | -69% | 1 | 1 | 0% | 1,026 | 1,154 | +12% | 0 | 0 | — |
case-17 | fail→pass | 26,446 | 4,055 | -85% | 1 | 1 | 0% | 3,975 | 1,087 | -73% | 0 | 0 | — |
case-18 | fail→pass | 13,004 | 1,495 | -89% | 1 | 1 | 0% | 1,953 | 978 | -50% | 0 | 0 | — |
case-19 | pass→pass | 8,732 | 2,642 | -70% | 1 | 1 | 0% | 1,239 | 1,229 | -1% | 0 | 0 | — |
case-20 | pass→pass | 3,124 | 2,737 | -12% | 1 | 1 | 0% | 440 | 1,202 | +173% | 0 | 0 | — |
case-21 | pass→pass | 6,615 | 3,412 | -48% | 1 | 1 | 0% | 1,018 | 1,317 | +29% | 0 | 0 | — |
case-22 | pass→pass | 7,948 | 3,856 | -51% | 1 | 1 | 0% | 1,141 | 1,420 | +24% | 0 | 0 | — |
case-23 | pass→pass | 11,400 | 8,905 | -22% | 1 | 1 | 0% | 1,731 | 2,114 | +22% | 0 | 0 | — |
DecimalAI ran this skill against gemini-3.6-flash twice over the same eval suite — once with the skill loaded and once without — and compared the two runs case by case. 23 cases were attempted. The headline lift of +65 percentage points is the difference between those two pass rates over the 23 comparable cases.
Without the skill loaded, the model failed this case. With it loaded, the same prompt on the same model passed. This is one improved case from the latest verified run; every case, including any that regressed, is in the table above.
| Model | Method | Date | Lift |
|---|---|---|---|
| gemini-3.6-flash | verified | 8/19/2026 | +61% |
Other measured skills in the registry, with their headline benchmark lift.