Skip to content

vercel-labs/agent-skills / skills/vercel-optimize

vercel-optimize

Verified

Use for Vercel cost and performance optimization on deployed projects, especially Next.js, SvelteKit, Nuxt, and limited Astro apps. Collect Vercel metrics, usage, project config, and code scan results first; investigate only metric-backed candidates; produce ranked recommendations grounded in verified files and version-aware Vercel/framework docs. Trigger for Vercel bill reduction, slow or expensive routes, caching opportunities, Function Invocations, Build Minutes, Fast Data Transfer, Core Web Vitals, Bot Management, Fluid compute, or cost breakdown requests.

FileKindSize
AGENTS.mdmarkdown1.9 KB
CONTRIBUTING.mdmarkdown2.2 KB
README.mdmarkdown4.4 KB
SKILL.mdskill md16.9 KB
lib/auth-route.mjsscript807 B
lib/budget-summary.mjsscript7.0 KB
lib/citations.mjsscript3.9 KB
lib/cost-coverage.mjsscript8.4 KB
lib/dedup-recs.mjsscript9.9 KB
lib/deep-dive.mjsscript10.2 KB
lib/display-labels.mjsscript7.4 KB
lib/extract-claims.mjsscript18.7 KB
lib/framework-support.mjsscript2.1 KB
lib/gates/build-minutes-fanout.mjsscript2.9 KB
lib/gates/cold-start.mjsscript2.7 KB
lib/gates/contract.mjsscript3.1 KB
lib/gates/cwv-poor.mjsscript3.3 KB
lib/gates/external-api-slow.mjsscript2.1 KB
lib/gates/hard-gates.mjsscript2.5 KB
lib/gates/index.mjsscript2.0 KB
lib/gates/isr-overrevalidation.mjsscript2.4 KB
lib/gates/middleware-heavy.mjsscript2.0 KB
lib/gates/observability-events-attribution.mjsscript2.2 KB
lib/gates/platform-bot-protection.mjsscript5.0 KB
lib/gates/platform-fluid-compute.mjsscript3.5 KB
lib/gates/region-misconfig.mjsscript3.0 KB
lib/gates/route-errors.mjsscript3.0 KB
lib/gates/scanner-driven.mjsscript5.1 KB
lib/gates/select-candidates.mjsscript4.6 KB
lib/gates/slow-route.mjsscript4.3 KB
lib/gates/types.d.tsscript893 B
lib/gates/uncached-route.mjsscript3.9 KB
lib/gates/usage-spike-triage.mjsscript5.0 KB
lib/grade-recommendation.mjsscript6.7 KB
lib/impact-label.mjsscript5.2 KB
lib/impact-magnitude.mjsscript2.1 KB
lib/investigation-brief.mjsscript34.3 KB
lib/observation-safety.mjsscript9.3 KB
lib/project-facts.mjsscript4.3 KB
lib/queries.mjsscript10.1 KB
lib/reconcile-candidates.mjsscript15.6 KB
lib/render-report.mjsscript38.2 KB
lib/repo-root.mjsscript3.1 KB
lib/route-normalize.mjsscript8.4 KB
lib/sanitizers/bot-protection-certainty.mjsscript1.9 KB
lib/sanitizers/cache-tag-invalidation-certainty.mjsscript1.5 KB
lib/sanitizers/count-correct.mjsscript1.9 KB
lib/sanitizers/function-duration-invocations.mjsscript1.4 KB
lib/sanitizers/index.mjsscript3.0 KB
lib/sanitizers/middleware-conflict.mjsscript1.5 KB
lib/sanitizers/missing-citation.mjsscript497 B
lib/sanitizers/pre-release.mjsscript2.2 KB
lib/sanitizers/rate-limit.mjsscript2.6 KB
lib/sanitizers/rendering-mode-mislabel.mjsscript1.6 KB
lib/sanitizers/undeclared-dep.mjsscript2.8 KB
lib/sanitizers/vercel-directive-strip.mjsscript1.4 KB
lib/sanitizers/window-units.mjsscript869 B
lib/scanners/cache-components-suspense-dedupe.mjsscript4.3 KB
lib/scanners/edge-heavy-import.mjsscript3.7 KB
lib/scanners/force-dynamic.mjsscript1.5 KB
lib/scanners/headers-in-page.mjsscript1.5 KB
lib/scanners/index.mjsscript1.4 KB
lib/scanners/large-static-asset.mjsscript3.2 KB
lib/scanners/max-age-without-s-maxage.mjsscript1.6 KB
lib/scanners/middleware-broad-matcher.mjsscript1.9 KB
lib/scanners/missing-cache-headers.mjsscript3.8 KB
lib/scanners/prisma-include-tree.mjsscript1.5 KB
lib/scanners/region-pin-in-config.mjsscript3.4 KB
lib/scanners/source-maps-production.mjsscript1.3 KB
lib/scanners/sveltekit-prerender-missing.mjsscript2.1 KB
lib/scanners/turbo-force-bypass.mjsscript4.9 KB
lib/scanners/unoptimized-image.mjsscript4.4 KB
lib/scanners/use-cache-date-stamp.mjsscript3.9 KB
lib/support-topics.mjsscript12.5 KB
lib/throttle.mjsscript8.8 KB
lib/util.mjsscript544 B
lib/vercel.mjsscript30.4 KB
lib/verify-claim.mjsscript59.6 KB
lib/workspace-resolver.mjsscript17.1 KB
metadata.jsonconfig903 B
references/candidates.mdmarkdown8.3 KB
references/data-collection.mdmarkdown20.4 KB
references/docs-library.jsonconfig30.8 KB
references/doctrine.mdmarkdown9.3 KB
references/observability-plus.mdmarkdown5.2 KB
references/playbooks/README.mdmarkdown2.7 KB
references/playbooks/ai-application.mdmarkdown3.8 KB
references/playbooks/api-service.mdmarkdown2.4 KB
references/playbooks/content-site.mdmarkdown2.4 KB
references/playbooks/ecommerce.mdmarkdown2.7 KB
references/playbooks/marketing.mdmarkdown2.6 KB
references/playbooks/saas.mdmarkdown2.8 KB
references/playbooks/sveltekit.mdmarkdown3.9 KB
references/recommendations.mdmarkdown10.5 KB
references/scanner-patterns.mdmarkdown13.4 KB
references/scoring.mdmarkdown12.2 KB
references/support-topics/README.mdmarkdown1.7 KB
references/support-topics/astro-edge-middleware-scope.mdmarkdown1012 B
references/support-topics/astro-output-mode-and-isr.mdmarkdown1.1 KB
references/support-topics/auth-preserving-parallelization.mdmarkdown1.2 KB
references/support-topics/bot-protection-product-guardrails.mdmarkdown1.1 KB
references/support-topics/build-minutes-monorepo-fanout.mdmarkdown1.2 KB
references/support-topics/cache-components-static-shell-boundaries.mdmarkdown1.1 KB
references/support-topics/cache-components-suspense-dedupe-pitfall.mdmarkdown1.3 KB
references/support-topics/cdn-cache-auth-safety.mdmarkdown1.0 KB
references/support-topics/cold-start-initialization-bundle.mdmarkdown1.1 KB
references/support-topics/core-web-vitals-client-bottlenecks.mdmarkdown1.0 KB
references/support-topics/database-egress-pooling-region.mdmarkdown1015 B
references/support-topics/dynamic-rendering-traps.mdmarkdown1.1 KB
references/support-topics/external-api-critical-path-platform.mdmarkdown1.1 KB
references/support-topics/external-api-critical-path.mdmarkdown978 B
references/support-topics/fast-data-transfer-payloads.mdmarkdown1.0 KB
references/support-topics/fluid-compute-caveats.mdmarkdown1.1 KB
references/support-topics/function-duration-io-and-after.mdmarkdown1.0 KB
references/support-topics/function-invocation-reduction.mdmarkdown1.2 KB
references/support-topics/function-region-misconfiguration-ttfb.mdmarkdown1.2 KB
references/support-topics/image-optimization-cost-control.mdmarkdown1.0 KB
references/support-topics/isr-revalidation-static-generation.mdmarkdown1.4 KB
references/support-topics/middleware-proxy-edge-cost.mdmarkdown991 B
references/support-topics/next-fetch-revalidate-floor.mdmarkdown1.0 KB
references/support-topics/next-font-cls-self-hosting.mdmarkdown906 B
references/support-topics/next-heavy-ui-lazy-load-boundaries.mdmarkdown1.0 KB
references/support-topics/next-image-lcp-preload-sizes.mdmarkdown1.0 KB
references/support-topics/next-route-handler-get-cache-defaults.mdmarkdown1.0 KB
references/support-topics/next-script-third-party-strategy.mdmarkdown1.0 KB
references/support-topics/nextjs-version-cache-semantics.mdmarkdown996 B
references/support-topics/not-found-catchall-request-waste.mdmarkdown1.2 KB
references/support-topics/nuxt-route-rules-cache-isr.mdmarkdown1.2 KB
references/support-topics/observability-events-cost-attribution.mdmarkdown1.1 KB
references/support-topics/post-response-work-waituntil.mdmarkdown1009 B
references/support-topics/route-error-durable-offload.mdmarkdown1.3 KB
references/support-topics/route-error-runtime-limits.mdmarkdown988 B
references/support-topics/runtime-cache-reusable-data.mdmarkdown1.2 KB
references/support-topics/sveltekit-isr-prerender-safety.mdmarkdown1.1 KB
references/support-topics/sveltekit-split-cold-start-tradeoff.mdmarkdown1.0 KB
references/support-topics/usage-spike-triage.mdmarkdown1018 B
references/support-topics/use-cache-date-stamp-isr-write-amplifier.mdmarkdown1.2 KB
references/support-topics/use-cache-remote-shared-origin-data.mdmarkdown1.1 KB
references/support-topics/workflow-resumable-stream-routes.mdmarkdown1.2 KB
references/verification.mdmarkdown8.8 KB
references/voice.mdmarkdown3.6 KB
scripts/budget-summary.mjsscript2.0 KB
scripts/build-docs.mjsexecutablescript3.3 KB
scripts/check-citations.mjsscript3.4 KB
scripts/check-docs-fresh.mjsexecutablescript3.8 KB
scripts/collect-signals.mjsexecutablescript21.1 KB
scripts/collect-sub-agent-outputs.mjsscript9.9 KB
scripts/deep-dive.mjsexecutablescript12.0 KB
scripts/gate-investigations.mjsexecutablescript6.2 KB
scripts/merge-signals.mjsscript6.2 KB
scripts/prepare-investigation-brief.mjsscript7.9 KB
scripts/reconcile-candidates.mjsscript2.4 KB
scripts/render-report.mjsscript17.0 KB
scripts/scan-codebase.mjsexecutablescript11.3 KB
scripts/verify-and-regen.mjsscript14.9 KB
scripts/verify-finding.mjsexecutablescript698 B

references/playbooks/ai-application.md

View on GitHub
# AI application

LLM-backed apps, agents, code-sandbox tools, RAG pipelines. Cost shape is dominated by per-token AI Gateway spend and Sandbox active-compute time, not edge requests or function duration. Many AI customers also have a SaaS surface (auth, dashboards), but the cost lever lives upstream of the dashboard.

## Typical billing shape

AI Gateway > Sandbox Active Compute > Function Duration > Function Invocations. Edge Requests usually quiet; ISR rarely applies. Observability Events can climb fast if every tool-call span is captured at full fidelity.

## Priority patterns

1. **Provider failover.** Configure AI Gateway with an active-active fallback chain across providers (OpenAI + Anthropic, or model-family pairs). Critical-path agents must not be single-provider — a 429 from one provider becomes a user-visible outage otherwise. Field example: MELI runs homegrown active-active routing because retry-on-error against a single provider degraded their NLP-on-support flow.
2. **OIDC keyless auth, not explicit API keys.** In production, use the AI Gateway OIDC binding so requests are signed by deployment identity. In local dev, `vercel env run -- <cmd>` rotates OIDC each run. An explicit `AI_GATEWAY_API_KEY` in repo env vars is a regression — it bypasses keyless and creates a long-lived secret.
3. **Sandbox reuse over per-request `Sandbox.create`.** Each fresh sandbox costs at least 1 minute of billed compute (boot + teardown rounded up). When isolation isn't required (single-tenant agents, shared workspaces), pool sandboxes by name (`sandbox.get(name)`) — auto-snapshot on death + auto-resume on next get is the persistence model.
4. **`after()` / `waitUntil()` for tool logging.** Tool-call telemetry, audit writes, and analytics should never block the user response. Use `after()` (Next 15+) or `waitUntil()` from `@vercel/functions` for any write that doesn't affect the streamed response.
5. **Fluid Compute for JIT/process warmth.** Streaming LLM responses benefit from warm processes; the GraphQL/Apollo JIT cache + persisted-document plans only pay back when processes survive across requests. Fluid is the default; disabling it on AI workloads is almost always wrong.

## Frequent gotchas

- **Single-provider lock-in.** "We're using AI Gateway" doesn't imply failover — the provider list still has to be configured. A single-provider gateway is a thinner wrapper, not multi-provider resilience.
- **Sandbox-per-request.** `new Sandbox(...)` inside a per-request handler with no `id` argument creates a fresh microVM each time. Cheaper to pool when isolation allows.
- **BYOK fallback cost invisible.** AI Gateway with BYOK silently falls back to system credits on 429 / provider outage; cost migrates from "free BYOK" to "billed credits" without a separate signal unless tracked.
- **Observability Events runaway.** Captured every tool call + every streamed delta at 100% sampling — events SKU climbs above 30% of bill. Cap span cardinality before scaling traffic.

## Cross-references

- [external-api-critical-path](../support-topics/external-api-critical-path.md) — sequential vs parallel calls; AI Gateway is one external API among others
- [fluid-compute-caveats](../support-topics/fluid-compute-caveats.md) — module-state hazards and shared-instance caveats
- [function-duration-io-and-after](../support-topics/function-duration-io-and-after.md) — `after()` for post-response tool logging
- [observability-events-cost-attribution](../support-topics/observability-events-cost-attribution.md) — when Observability Events climb above 20% of bill
- [use-cache-remote-shared-origin-data](../support-topics/use-cache-remote-shared-origin-data.md) — caching shared LLM context or embedding lookups
- `https://vercel.com/docs/ai-gateway` — provider configuration, failover chain
- `https://vercel.com/docs/vercel-sandbox` — `sandbox.get(name)` and active-compute billing
writing-skillsobra/superpowersUse when creating new skills, editing existing skills, or verifying skills work before deploymentStars: 322kUpdated Claude CodeCodexVerifiedInstall
xlsxanthropics/skillsUse this skill any time a spreadsheet file is the primary input or output. This means any task where the user wants to: open, read, edit, or fix an existing .xlsx, .xlsm, .xltx, .csv, or .tsv file (e.g., adding columns, computing formulas, formatting, charting, cleaning messy data); create a new spreadsheet from scratch or from other data sources; or convert between tabular file formats. Trigger especially when the user references a spreadsheet file by name or path — even casually (like "the xlsx in my downloads") — and wants something done to it or produced from it. Also trigger for cleaning or restructuring messy tabular data files (malformed rows, misplaced headers, junk data) into proper spreadsheets. The deliverable must be a spreadsheet file. Do NOT trigger when the primary deliverable is a Word document, HTML report, standalone Python script, database pipeline, or Google Sheets API integration, even if tabular data is involved.Stars: 198kUpdated Claude CodeCodexVerifiedInstall
shipping-and-launchaddyosmani/agent-skillsPrepares production launches. Use when preparing to deploy to production, or when asking what needs to be in place before shipping. Use when you need a pre-launch checklist, when setting up monitoring, when planning a staged rollout, or when you need a rollback strategy.Stars: 119kUpdated Claude CodeCodexVerifiedInstall
childhoodthedotmack/claude-memBabysit a shipped release after merge, publication, or deployment. Verify published packages, production behavior, installed plugins, and late CI or user reports until the observation window is clean; use babysit for PR readiness before release.Stars: 118kUpdated Claude CodeCodexVerifiedInstall