2026 06 16 Skill Inventory
Cloudflare skill-guided implementation inventory — 2026-06-16
Scope: read-only inventory of Cloudflare-related implementation, configuration, and documentation surfaces in this repository. No Cloudflare APIs were called, no DNS was mutated, no deploys were run, and no secrets were printed or added.
Skill guidance applied#
Reviewed before inventory:
/home/vchavkov/.config/brain/notes/skills/cloudflare/INDEX.md/home/vchavkov/.config/brain/.agents/skills/cloudflare-pages/SKILL.md/home/vchavkov/.config/brain/.agents/skills/cloudflare-workers/SKILL.md/home/vchavkov/.config/brain/.agents/skills/cloudflare-storage/SKILL.md/home/vchavkov/.config/brain/.agents/skills/cloudflare-dns/SKILL.md/home/vchavkov/.config/brain/.agents/_pi/skills/service-registry/SKILL.md
Operational constraints carried into this audit:
- Prefer Wrangler for Pages/Workers deploy surfaces; use REST only where Wrangler lacks coverage.
- Treat KV/D1/R2 as explicit storage bindings, with destructive operations and token scopes handled carefully.
- For DNS/custom domains, coordinate Cloudflare-hosted zones with the Cloudflare DNS guidance and registry-owned
svc.*/Linode/split-horizon hostnames with the Service Registry guidance. - Do not hardcode or expose Cloudflare tokens; use
CLOUDFLARE_API_TOKEN/CF_API_TOKENand account/zone IDs only through environment/configured deploy contexts.
Evidence commands#
Read-only discovery commands used:
bash
1find . \( -path '*/node_modules/*' -o -path './.git/*' -o -path '*/.turbo/*' -o -path '*/.next/*' -o -path '*/.next-build/*' -o -path '*/.next-dev/*' -o -path '*/.open-next/*' -o -path '*/.wrangler/*' -o -path '*/dist/*' -o -path '*/out/*' -o -path '*/build/*' -o -path '*/coverage/*' -o -path './.worktrees/*' -o -path '*/content/courses/*' -o -path '*/content/learnings/*' -o -path '*/content/presentations/*' -o -path '*/_blog/*' \) -prune -o \( -iname '*wrangler*' -o -iname '*cloudflare*' -o -iname '*opennext*' \) -print | sort23rg -l -i "cloudflare|wrangler|opennext|pages\.dev|CF_|CLOUDFLARE|kv_namespaces|d1_databases|r2_buckets|@cloudflare|workers" \4 --glob '!**/node_modules/**' --glob '!.git/**' --glob '!**/.turbo/**' \5 --glob '!**/.next/**' --glob '!**/.next-build/**' --glob '!**/.next-dev/**' \6 --glob '!**/.open-next/**' --glob '!**/.wrangler/**' --glob '!**/dist/**' \7 --glob '!**/out/**' --glob '!**/build/**' --glob '!**/coverage/**' \8 --glob '!**/content/courses/**' --glob '!**/content/learnings/**' --glob '!**/content/presentations/**' \9 --glob '!**/_blog/**' --glob '!**/public/img/**' --glob '!**/public/images/**' \10 --glob '!**/spec/**' | sort1112find . \( -path '*/node_modules/*' -o -path './.git/*' -o -path '*/.turbo/*' -o -path '*/.next/*' -o -path '*/.next-build/*' -o -path '*/.next-dev/*' -o -path '*/.open-next/*' -o -path '*/.wrangler/*' -o -path './.worktrees/*' \) -prune -o -name 'wrangler.*' -print | sortGenerated/local artifacts intentionally excluded from the inventory include node_modules/, .next*/, .open-next/, .wrangler/, dist/, out/, build/, coverage/, .turbo/, and .worktrees/. Course/blog content hits were also excluded where they were educational content rather than repo implementation or operator surfaces.
Primary deploy/config inventory#
App deployment surfaces#
Storage/search surfaces#
Cache/performance/quality surfaces#
Documentation surfaces#
DNS and registry-relevant findings#
apps/docs/wrangler.jsoncis the only inspected app config with explicit Workerroutesunderassistance.bg; future route changes should follow Cloudflare Workers routing guidance and coordinate with DNS/custom-domain ownership..docs/cloudflare-deployment.mdstateswww.assistance.bgis production Cloudflare DNS, while dev/prod subzones are delegated to Linode and should be managed through the registry MCP/API.- Service Registry skill guidance applies whenever new
svc.assistance.bg,svc.chavkov.com, split-horizon, Caddy, or registry-owned DNS/Caddy exposure work is proposed. This audit did not call the registry API. - Cloudflare DNS skill guidance applies to Cloudflare-hosted public records: resolve zone IDs by name, use idempotent upserts, dry-run/confirm before mutation, and avoid hardcoded zone IDs in operator snippets.
Not included as implementation surfaces#
- Generated outputs:
.open-next/,.wrangler/,.next*, appbuild/,dist/, andout/directories. - Installed dependencies under
node_modules/, including nested Wrangler/OpenNext packages. - Course lessons, blog articles, and image/icon references that mention Cloudflare as educational/product content rather than deployment/configuration implementation.
- Private runtime state or account details; no Cloudflare account/resource queries were made.