agentleFS
awaikawaik/direct-mcp-ai-project/.claude/skills/lidfly-site-commerce/SKILL.md

lidfly-site-commerce

Skill20 starsChanged 4 days agoMITView on GitHub

Создавать, оформлять и проверять сайты LidFly через MCP v3: визуальная концепция, шаблоны, блоки, изображения, адаптивность и visual QA. Также страницы, базы знаний, SEO/GEO, feeds, файлы, лиды, публикация, товары, остатки, заказы и платежи с точным scope и защитой секретов YooKassa.

The file

We split this skill into the rules it gives an agent

6 rules. About 0 are common, 2 are distinctive, and 4 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈2

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 4

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Very specific:

    Read site/references and lidfly_get_site_theme.design_system: v2 owns rhythm/type.

    line 22
  2. Distinctive:

    Read Design system, Creative direction.

    line 24
  3. Distinctive:

    If needed, use Visual assets: roles/crops, inspected images/bindings.

    line 28
  4. Very specific:

    Follow Visual QA: view desktop/mobile screenshots, test/fix/recheck interactions and revision.

    line 30
  5. Very specific:

    Use native site capability, never CSS/HTML/custom fields/Bitrix24.

    line 76
  6. Very specific:

    No webhook needed; widget is independent of server forms.

    line 93

≈ “Common” is an estimate: at least 1 in 5 of the 372,429 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

LidFly Site Commerce

Contracts

«Тема оформления» means visual tokens (colors/fonts/radius/renderer choices). «Шаблон сайта» means persistent site-level design: header/footer, product page/card, checkout and page blueprints. Generated storefront pages are not «templates» without design_template_id. Store/provider tools backed by PostgreSQL own Commerce state; /sites HTML is an artifact. Never show, echo or save seller YooKassa secrets in visible documents.

Workflow

Visual work

Creation/redesign cycle within scope:

  1. Read site/references and lidfly_get_site_theme.design_system: v2 owns rhythm/type. Define audience, hierarchy, palette, grid, spacing, image treatment. Read Design system, Creative direction.
  2. Map native blocks/mobile order using Frontend craft. Blueprints/compositions are starting points, not quotas; knowledge pages need readable structure, not forced sales heroes/oversized cards.
  3. If needed, use Visual assets: roles/crops, inspected images/bindings. Preserve content, invent no evidence or paid requirement.
  4. Follow Visual QA: view desktop/mobile screenshots, test/fix/recheck interactions and revision. Technical pass is not visual approval.

Corrections preserve direction; order/inventory/SEO edits need no redesign. Load references by stage via get_skill_resource(name="lidfly-site-commerce", path=...) or local files. Unavailable viewing means incomplete QA.

Native operations

Choose whole-card navigation/zoom/CTA/form explicitly. Use section snapshots for corrections, blueprints for composition. Filter definitions by purpose/description/ visible_when; paginate while has_more. Explain unsupported behavior; a button cannot replace whole-card navigation. Disclose warnings and test clicks, anchors, unwanted zoom, keyboard and mobile.

For design_template_id="knowledge-base", use $lidfly-knowledge-maintainer for ingest, query-to-wiki, provenance, relations, findings, changesets and lint; never emulate knowledge updates with sequential page writes.

  1. Resolve site/store/owner/project with get_provider_context({provider:"lidfly",query?}); use returned scope. lidfly_list_sites only when no URL/id/domain/name is known.
  2. search_tools → first-use get_tool_schema. call_tool reads; call_write_tool publishes/uploads, changes store/order/payment/CRM policy, resolves delivery and generates images. Paid images need explicit prompt, format and crop confirmation.
  3. Serialize same-site writes with fresh publication_revision and publication_write.status="idle". If busy, await, reread scope and verify prior write before at most one retry.

New/substantially revised managed sites follow Methodology: page/image scope → compact snapshots → blueprint → acceptance-bound changeset → exact-revision desktop/mobile checks. Readiness requires all promised routes, assets and QA, with warnings disclosed. Save desired-state edits sequentially; publish once per site.

Focused corrections

Preserve target neighbors/identities. No blueprint, reference composition or full inventory unless composition changes. Follow Visual QA: read/patch each taxonomy node (patch.slotOperations, op="update_block"), preview summary, publish once, inspect exact revision and returned block index. That reference defines desktop/mobile crops, conditional tablet checks and next_safe_call recovery.

Privacy consent

Use native site capability, never CSS/HTML/custom fields/Bitrix24. Read lidfly_get_site_privacy_consent, publish internal documents, then lidfly_update_site_privacy_consent with exact revisions; reread. form.required=true: one required unchecked checkbox, distinct consent_url/privacy_policy_url. Changed enabled text/links need a new version. analytics.required=true: accept/reject banner blocks Metrika/optional tracking; new versions ask again. Verify lidfly_get_leads.consent: time/version, immutable text, both document URLs and page URL. Visual presence is insufficient.

Bitrix24 / B242YA widget

Pass full widget_code, optional b242ya_code, exact subdomain to lidfly_prepare_bitrix24_widget; never extract URLs or replace blocks. Show portal/origins/components/scope/conflicts. Only ready=true: apply agreed normalized_arguments with lidfly_apply_bitrix24_widget. Conflicts need fresh preparation/agreement, never CAS substitution. disable retains settings; reset removes them. Pending/lost response: read operation journal. No webhook needed; widget is independent of server forms. Installation proves no ClientID delivery. Own t.me links without a second button: mode=telegram_links, one telegram_bot, B242YA and CRM tracking required. Real CRM/Telegram inquiries need separate authorization. Sources: Bitrix24, B242YA, checked 08.10.2026.

Progressive References

Load by task/stage; visual guides are mandatory for redesign, proportional for corrections.

  • Design system v2: styles/composition.
  • Managed pages: widgets/metadata/video.
  • Site chrome: header/footer/templates.
  • Static sites: archive preview/deploy.
  • Commerce: products/imports/add-ons/feeds.
  • Galleries: albums/preview/apply/Bitrix.
  • SEO/feeds: GEO/Organization/articles/RSS.
  • Methodology: initial chat / legacy get_methodology.

Workspace

Resolve workspace_project_id before saving business/client decisions, documents or scheduled follow-ups.

Client materials and evidence

Stories are editorial content, not testimonials. Invent no ratings/stars, counts, results, certificates or quotes. Use supplied facts; record file/message/URL provenance in project notes. Label edits, never call them verbatim. List missing photos/testimonials/proof; preserve other blocks, asset paths and design.

New page images/logos: one lidfly_preview_site_changeset → lidfly_apply_site_changeset with import_image (key + chat_attachment_id or image_url), upsert_page, @asset:key. Confirm normalized arguments, digest and effects together. Standalone libraries may use separate uploads. patch_theme, set_site_chrome, set_site_css own theme, chrome and CSS. Never substitute page themePreset for site_style; if absent in changesets, read its tool schema and prepare a separate agreed action. Missing legacy canonical_path/public_url: reread that site's lidfly_list_assets, never infer paths from filenames.

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.