What the origin notices mean
An origin notice identifies an article, a connected explanatory section or an independently usable output. On pages with mixed content, it applies to the named section. It confirms neither the accuracy of the statements nor security, legal compliance or subject-matter approval.
Not every page or interface instruction has its own AI notice. This does not mean that text without a notice was written entirely by people or reviewed by a person. The origin of an individually sent message cannot be inferred from a notice on a linked web page.
How text and illustrations are created
We use language models and research and coding agents to create and revise content. The proprietor selects topics and tasks, examines the results and decides whether to publish them. This examination is not subject-matter approval of every statement.
AI tools are also used to create design, software and technical diagrams. Schematic diagrams explain relationships. They do not establish that a particular system has been measured. Synthetic product examples use artificial example data and are labelled as such.
Measurement, assessment and explanation
The public Web, Email and API Checks collect technical observations, for example by retrieving responses, querying DNS or testing connections within the stated scope. A measured value belongs to the specified measurement time. Features that were not measured are neither passed nor failed checks.
Check catalogues and fixed rules classify observations and determine points and grades. These assessments are our product methodology, not a comprehensive security judgement. An overall grade may be unavailable if the measurement basis is insufficient. The prewritten explanations and recommendations were created using AI and are assigned to findings. In the assessed process, a language model does not write them afresh for each individual check.
A technically correct measurement does not automatically make its associated recommendation suitable for every environment. Measurement errors, incomplete signals and incorrect or unsuitable explanations remain possible. The limits of a check and unresolved findings are therefore part of the result.
A public report downloaded again may be generated from the stored measurement using a newer presentation. This does not perform a new measurement. The measurement time, measurement catalogue used and document generation are distinct details.
Questionnaires evaluate the information you supply using fixed rules. They are not technical measurements or proof of actual implementation. Synthetic samples use example data to illustrate a structure or situation without having performed the measurement shown.
Which automated checks take place
Automated text checks look for specific problematic wording in defined content areas, such as unsupported security or compliance promises. Technical checks examine selected functions, data processing and presentation. A passing test or build confirms only the technical subject tested; it does not establish the subject-matter accuracy of every sentence.
During selected revisions, AI agents have compared statements with documented primary sources and corrected them accordingly, for example in NIS2 content. These source comparisons go beyond spelling or wording checks. Their scope differs by content; they do not constitute a complete source review of all content. AI-assisted source comparisons and further model reviews can also contain errors.
Comparing a direct quotation can check whether it matches a source. Whether the source is current, relevant to the particular case and correctly interpreted are separate questions. Review results are not applied indiscriminately to other articles or reports.
Limits of human review
There is no comprehensive human subject-matter review of each item of content. Selecting topics, examining results and publishing them are not equivalent to subject-matter approval of every statement. Automated tests, research agents and additional AI models do not replace such human review.
For this AI-created content, we do not rely on the exemption for human review or editorial control under the second subparagraph of Article 50(4) of the EU AI Act. Whether an origin disclosure is required also depends first on the content, purpose and publication. A notice does not remove other applicable obligations.
What dates mean
A stated source or revision date identifies the documented work and its review scope. It is neither human approval nor a promise of continuing currency. Not all content has the same review history or a stated source date.
A scan date refers to the technical measurement. A file’s generation date refers to its compilation. A new build or visual relaunch is not, by itself, a new fact check. Laws, standards and provider interfaces may change after a revision.
Before relying on material statements, check them against the cited sources and your specific environment.
Reporting possible errors
You can report possible errors to info@wolf-agents.com. Identify the page or output and the specific passage. Do not include confidential report data or access credentials.
The operator identified in the legal notice is responsible for publication.
English summary: how these results are created
Wolf-Agents uses AI tools to create and revise explanatory content, design and software. An AI disclosure identifies the relevant article, explanatory section or document. It is a statement about origin, not a guarantee of accuracy, security or compliance. The absence of a label does not mean that a text was written or professionally reviewed by a human.
The public scanners collect technical observations within the stated scope. Scores and grades follow fixed product rules. Prewritten explanations created using AI are assigned to the findings. Missing measurements do not count as passed checks, and an overall grade may be unavailable. A new download can use a newer presentation of a saved measurement; it does not perform a new scan. Synthetic examples use example data, and questionnaires evaluate the information supplied by the user.
Automated checks examine selected wording, functions and presentation. During selected revisions, AI agents have also compared statements with documented primary sources and corrected content. This does not amount to verification of every statement across all content. A quote match, a successful build or a further model review is not a human subject-matter review.
There is no comprehensive human subject-matter review of each item. We do not rely on the human review or editorial control exemption in Article 50(4) of the EU AI Act for these AI-created contents. Source and revision dates describe the documented work concerned; they do not promise continuous accuracy or updating. A disclosure does not remove other applicable obligations.
Possible errors may be reported to info@wolf-agents.com, identifying the page or document and the relevant passage. Please do not send confidential report data or access credentials. The provider responsible for publication is identified in the legal notice.