Open research · 2026 edition
State of Agent Skills 2026.
A transparent look at what is actually inside an agent-skill registry: which records look like reusable skills, which are maintained, which declare a license, what can be installed, and how little real outcome evidence the ecosystem still has.
20,466
Indexed candidates
985
Analyzed sample
7%
Strict agent skills
1
Agent Proven
Executive finding
Popularity is not readiness.
GitHub adoption is useful, but it does not prove task fit, safe installation, or a successful agent run. This report keeps those signals separate.
- 017.2% of the analyzed sample has strict Agent Skill evidence in its repository name, README, description, or topics.
- 0212.2% describes an agent workflow or agent product without strict reusable-skill evidence.
- 0329.8% is better described as a domain workflow tool or ecosystem dependency than a strict agent skill.
- 041.9% is a list or collection rather than one installable skill.
- 0599.9% was pushed within 180 days of report generation.
- 0681% has a declared license that is not marked unknown.
- 07100% reaches the current 72-point high-trust threshold in this quality-ranked sample; Trust remains a metadata signal, not proof of task success.
- 081 analyzed skill has at least one reported agent outcome, covering 1 outcome in total.
Readiness signals
What the analyzed sample can support.
- Maintained within 180 days
- 984100%
- Known license
- 79881%
- Install path available
- 985100%
- 500+ GitHub stars
- 985100%
- 72+ Trust Score
- 985100%
- Safe-to-try audit
- 95597%
Domain coverage
Where the candidate supply is concentrated.
Track assignment uses category, task, use-case, and repository text signals. It is a discovery aid, not a claim that every record is production ready.
| Track | Candidates | Share | Maintained | Known license | High trust |
|---|---|---|---|---|---|
| Coding and developer agents | 595 | 60.4% | 594 | 485 | 595 |
| Research and knowledge work | 148 | 15% | 148 | 122 | 148 |
| Data, BI, and analytics | 103 | 10.5% | 103 | 82 | 103 |
| Design and creative production | 51 | 5.2% | 51 | 41 | 51 |
| Legal, policy, and compliance | 37 | 3.8% | 37 | 28 | 37 |
| Finance and quant workflows | 28 | 2.8% | 28 | 24 | 28 |
| Marketing and growth automation | 12 | 1.2% | 12 | 8 | 12 |
| Education and tutoring | 7 | 0.7% | 7 | 4 | 7 |
| Presentation and deck workflows | 3 | 0.3% | 3 | 3 | 3 |
| Football and World Cup analytics | 1 | 0.1% | 1 | 1 | 1 |
Evidence shortlist
Leading strict skill candidates.
Ordered by reported outcomes, Trust Score, skill-likeness, and GitHub adoption. This is a shortlist for evaluation, not an automatic-install endorsement.
| Skill | Track | Stars | Trust | Audit | Risk | Outcomes |
|---|---|---|---|---|---|---|
| Crawl4AI unclecode/crawl4ai | Research and knowledge work | 72K | 93 | 95 | Safe to try | 1 |
| Taste Skill Leonxlnx/taste-skill | Coding and developer agents | 61K | 95 | 96 | Safe to try | 0 |
| Code Review mattpocock/skills | Coding and developer agents | 164K | 94 | 95 | Safe to try | 0 |
| Open Design nexu-io/open-design | Coding and developer agents | 77K | 94 | 95 | Safe to try | 0 |
| Marketingskills coreyhaines31/marketingskills | Coding and developer agents | 37K | 94 | 95 | Safe to try | 0 |
| Nanoclaw nanocoai/nanoclaw | Coding and developer agents | 30K | 94 | 96 | Safe to try | 0 |
| Last30days Skill mvanhorn/last30days-skill | Research and knowledge work | 51K | 93 | 95 | Safe to try | 0 |
| Scientific Agent Skills K-Dense-AI/scientific-agent-skills | Data, BI, and analytics | 31K | 93 | 95 | Safe to try | 0 |
| Eigent eigent-ai/eigent | Coding and developer agents | 15K | 93 | 95 | Safe to try | 0 |
| Ponytail DietrichGebert/ponytail | Coding and developer agents | 79K | 93 | 95 | Safe to try | 0 |
| Cherry Studio CherryHQ/cherry-studio | Coding and developer agents | 48K | 93 | 95 | Safe to try | 0 |
| OpenMontage calesthio/OpenMontage | Coding and developer agents | 36K | 93 | 95 | Safe to try | 0 |
| Apple Design emilkowalski/skills | Design and creative production | 7.9K | 93 | 95 | Safe to try | 0 |
| SkillSpector NVIDIA/SkillSpector | Coding and developer agents | 13K | 93 | 95 | Safe to try | 0 |
| Pm Skills phuryn/pm-skills | Coding and developer agents | 22K | 93 | 95 | Safe to try | 0 |
| Huashu Design alchaincyf/huashu-design | Coding and developer agents | 19K | 92 | 93 | Safe to try | 0 |
| Guizang Ppt Skill op7418/guizang-ppt-skill | Presentation and deck workflows | 17K | 92 | 93 | Safe to try | 0 |
| Claude Seo AgriciDaniel/claude-seo | Coding and developer agents | 8.9K | 92 | 95 | Safe to try | 0 |
| Review Animations emilkowalski/skills | Coding and developer agents | 7.9K | 92 | 95 | Safe to try | 0 |
| Animation Vocabulary emilkowalski/skills | Design and creative production | 7.9K | 92 | 95 | Safe to try | 0 |
| Visual Explainer nicobailon/visual-explainer | Presentation and deck workflows | 9.2K | 92 | 95 | Safe to try | 0 |
| Emil Design Engineering emilkowalski/skills | Coding and developer agents | 7.9K | 92 | 95 | Safe to try | 0 |
| Cognee topoteretes/cognee | Research and knowledge work | 28K | 92 | 94 | Safe to try | 0 |
| Frontend Slides zarazhangrui/frontend-slides | Coding and developer agents | 22K | 92 | 95 | Safe to try | 0 |
| Ui Ux Pro Max Skill nextlevelbuilder/ui-ux-pro-max-skill | Coding and developer agents | 104K | 92 | 94 | Safe to try | 0 |
Methodology
Readable before repeatable.
Published Jul 10, 2026. Data refreshed Jul 10, 2026.
Population and sample
All public records with ai_review_approved=true in the OpenAgentSkill registry. MCP-only records are excluded by the registry read layer.
The highest-ranked 985 approved candidates by the current quality ordering. The sample is a product audit, not a random statistical sample.
Inclusion and measurement
- •Public GitHub repository or repository URL is available
- •Registry AI review is approved
- •MCP-only server records are excluded
- •Skill-likeness is measured separately from repository popularity
- •Trust and audit scores use public repository metadata and declared install information
Limitations
- •GitHub stars measure repository adoption, not agent task success
- •Candidate listings are not equivalent to verified maintainer claims
- •A public metadata audit is not a substitute for source review or sandbox execution
- •Agent Proven evidence remains sparse until agents report real outcomes
- •Repository metadata can change after this report is generated
Cite this report
OpenAgentSkill. “State of Agent Skills 2026.” Published July 10, 2026. Dataset and methodology available at https://www.openagentskill.com/reports/state-of-agent-skills-2026.