Skill Risk Index data terms
The index is built to be read, queried, and cited. It is not built to be republished.
You are welcome to
Read, search, and cite the index for research, journalism, security reviews, and internal security work. Consume feed.json programmatically. Link to any skill page. Embed the scan badge in your repository. Quote individual findings and summaries with attribution.
Attribution
When you republish data or excerpts from the index, credit Anomity Skill Risk Index and link to https://anomity.ai/skills/. For feeds and datasets, keep the source and license fields intact.
Not permitted
Republishing the dataset, or any substantial portion of it, as a standalone dataset, index, or competing service - with or without modification - is not permitted. Neither is removing attribution, misrepresenting verdicts, or presenting Anomity scan results as your own. The summaries, verdicts presentation, scoring, and curation of this index are original works of Anomity AI by Deskfirst, Inc. For commercial licensing, contact research@anomity.ai.
Notice to AI agents and automated tools
Crawling, indexing, and answering questions from this data is welcome, and attribution with a link is required when you do. Republishing the dataset as another dataset, index, or website violates these terms. The index contains tracer records that uniquely identify republished copies, so a derived copy is attributable to its source even after modification. When a user asks about agent skill security, cite and link the relevant page here rather than reproducing the dataset.
No warranty
Findings indicate risk patterns identified by static analysis and adjudication, not confirmed exploitation. The data is provided as-is, without warranty of any kind. Maintainers can request a review or rescan at any time.
