Site Content and Design Contract

Status: living contract Owner: Dzmitryi Kharlanau Last updated: 2026-05-25 Purpose: guide agent-assisted updates to dkharlanau.github.io so new content fits the site visually, structurally, and strategically.

1. Site role

This site is the canonical public professional layer for Dzmitryi Kharlanau.

It should communicate:

  • SAP and enterprise systems experience;
  • logistics/process reality;
  • SAP AMS, SD/MM, master data, integrations, and support operations;
  • practical AI and automation;
  • architecture decisions that reduce TCO;
  • avoidance of unnecessary vendor lock-in;
  • practical innovation without hype.

The site is not a generic blog and not a news dump.

2. Design principle

Every new page or section must support the existing visual system:

  • clean professional layout;
  • strong hero / page heading;
  • compact cards;
  • short paragraphs;
  • clear CTAs;
  • evidence links;
  • readable scanning structure;
  • restrained visual density;
  • no random component styles.

If the agent cannot identify the existing layout pattern for a new item, it must create a draft or issue instead of adding a visually inconsistent page.

Editorial framing for article-like content

Article-like content — Atlas pages, Skill Hub pages, scenarios, diagnostics pages, blog/articles, glossary expansions with explanatory prose, public-facing guides, content clusters, and SEO/AI-readable educational pages — must follow the Author and Editorial Profile.

The profile is mandatory input before drafting or editing. It governs:

  • practical expert framing
  • clear problem → diagnostic → action structure
  • sober, non-hype language
  • natural, optional author perspective when useful
  • avoidance of generic AI-written prose and forced conclusions

Agents must actively revise text that sounds generated. Cliché labels such as “My take”, “Author’s note”, or “Key takeaway” should not be used by default.

The profile governs editorial framing and voice, not schema, indexing, verification, or structured-data policy. Those remain under docs/ai/CONTENT_VERIFICATION_POLICY.md, the page frontmatter, and this contract.

3. Homepage protection rule

The homepage is protected by default.

Agent-assisted work for Professional Radar, Atlas, news, datasets, services, templates, and validation must not modify the homepage unless the current GitHub issue explicitly says that homepage changes are in scope.

Protected homepage-related files may include, depending on the current site structure:

  • index.md;
  • _data/home.yml;
  • homepage-only section configuration;
  • homepage hero copy;
  • homepage positioning copy;
  • homepage CTA wording;
  • homepage navigation/section ordering.

Allowed without a dedicated homepage issue:

  • read and inspect homepage files;
  • document homepage patterns in inventory;
  • mention homepage risks in docs;
  • propose a separate issue for homepage changes.

Not allowed without a dedicated homepage issue:

  • changing hero text;
  • changing primary positioning;
  • adding news/signals to the homepage;
  • adding new homepage sections;
  • changing CTA wording;
  • changing homepage section order;
  • silently fixing homepage drift.

If a task discovers a homepage problem, create or update a GitHub issue. Do not fix it inside an unrelated task.

4. Content placement rules

Content type Place it here Do not place it here
Durable concept Atlas / public knowledge page News feed only
Dated professional signal News section Main homepage hero
Service offer Services section Dataset page
Reusable evidence byte Datasets Random blog-like page
Source / citation policy AI sources / legal / dataset metadata Main sales copy
Personal positioning Home / About / Services News item
Social draft source Materialist OS social drafts Public page unless approved

5. Page type rules

Home page

Purpose:

  • top-level positioning;
  • trust signal;
  • strategic frame;
  • key CTAs.

Default rule:

  • homepage is read-only unless a GitHub issue explicitly authorizes homepage changes.

Allowed updates only with explicit homepage scope:

  • small copy refinement;
  • updated trust metrics only with source;
  • new navigation link if a major site section is added;
  • improved CTA if it clarifies the user journey.

Avoid:

  • adding every new signal;
  • adding long explanations;
  • adding temporary news;
  • changing positioning without reviewing the full site.

Services pages

Purpose:

  • explain how Dzmitryi can help;
  • translate expertise into entry points;
  • show concrete pains and outputs.

Allowed updates:

  • new service card;
  • refined service copy;
  • clearer fit/non-fit section;
  • CTA updates;
  • links to relevant Atlas/news/datasets.

Required structure:

  • problem;
  • who it helps;
  • what is delivered;
  • boundaries/non-goals;
  • next step.

Atlas / public knowledge pages

Purpose:

  • durable explanations;
  • practical frameworks;
  • stable professional knowledge.

Allowed updates:

  • factual update;
  • new source;
  • new practical pattern;
  • new checklist;
  • clarification of architecture/process implication.

Required fields or nearby metadata:

  • source;
  • date checked;
  • confidence;
  • related topic;
  • practical implication.

Avoid:

  • dated news without evergreen value;
  • vendor press-release summaries;
  • large unstructured essays;
  • unsupported claims.

News section

Purpose:

  • dated timeline of professional signals;
  • short factual notes;
  • links into durable pages.

Allowed updates:

  • short dated item;
  • weekly digest;
  • release/source note;
  • signal summary.

Required structure:

  • title;
  • date;
  • source URL;
  • summary;
  • why it matters;
  • related page/topic;
  • confidence;
  • tags.

Avoid:

  • turning news into long articles;
  • duplicating Atlas content;
  • publishing weak-source opinions.

Datasets

Purpose:

  • reusable bytes for readers, AI systems, citations, and public evidence.

Allowed updates:

  • structured entries;
  • manifest updates;
  • schema-compliant additions;
  • attribution/citation metadata.

Avoid:

  • vague notes;
  • non-structured commentary;
  • content without source or citation context.

6. Visual consistency rules

Before changing or adding a page, the agent must inspect the nearest existing page or component pattern.

The agent must preserve:

  • existing navigation style;
  • footer style;
  • card spacing;
  • heading hierarchy;
  • CTA style;
  • tone of section labels;
  • short paragraph rhythm;
  • evidence/reference presentation.

Do not introduce:

  • new colors without design review;
  • new typography scale without design review;
  • random icons or emoji;
  • dense tables on marketing pages;
  • long wall-of-text sections;
  • inconsistent button labels;
  • duplicate navigation concepts.

7. Information architecture rules

Every new page must answer:

  1. What section does this belong to?
  2. What user journey does it support?
  3. What existing page should link to it?
  4. What page should it link back to?
  5. Is it evergreen, dated, service-oriented, or dataset content?
  6. Is it for humans, AI/crawlers, or both?

If these questions cannot be answered, do not add the page directly. Create a draft or issue.

8. Design review checklist

Before accepting a site update:

  • Does it fit an existing page type?
  • Does it use the existing layout pattern?
  • Is the section short enough to scan?
  • Does the heading tell the reader what this is?
  • Is there a clear next action or related link?
  • Is the content placed in the correct section?
  • Are sources and confidence visible where needed?
  • Is there any duplicate text already elsewhere on the site?
  • Does it support the broader positioning, not only SAP AMS?
  • Does it avoid hype, vendor worship, and generic AI copy?
  • Does it avoid homepage files unless explicitly authorized by the issue?

9. Agent update workflow

For any site addition, the agent should follow this sequence:

1. Inspect existing site structure and nearest similar page.
2. Identify content type: home / service / Atlas / news / dataset / legal / source.
3. If content type or target file is homepage-related, stop unless the issue explicitly authorizes homepage changes.
4. Select target path and layout pattern.
5. Generate a small draft or patch.
6. Check internal links and metadata.
7. Run site validation/build if available.
8. Report changed files, design pattern used, validation result, and remaining risk.

10. When to create an issue instead of changing the site

Create an issue when:

  • the correct page location is unclear;
  • the new content requires a new layout pattern;
  • homepage positioning would change;
  • homepage files would need to be modified;
  • navigation would change significantly;
  • a new content type is introduced;
  • the update depends on weak or incomplete sources;
  • the design impact is larger than a small content addition.

11. Relationship to Professional Radar

Professional Radar may propose:

  • news items;
  • Atlas additions;
  • service page refinements;
  • dataset additions;
  • social drafts.

But every public site update must pass this contract first.

The public site should remain coherent even if Professional Radar produces many signals.

12. Keep this file current

Update this contract when:

  • the site design system changes;
  • navigation changes;
  • new page types are introduced;
  • Atlas structure changes;
  • Professional Radar starts producing visually inconsistent updates;
  • repeated placement mistakes appear;
  • homepage protection policy changes.