Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
npx skills add sanity-io/agent-toolkit --skill sanity-form-component-suspensegit clone --depth 1 https://github.com/sanity-io/agent-toolkitWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/skills/sanity-io/agent-toolkit/sanity-form-component-suspense)<a href="https://agentmods.dev/skills/sanity-io/agent-toolkit/sanity-form-component-suspense"><img src="https://agentmods.dev/badge/skills/sanity-io/agent-toolkit/sanity-form-component-suspense/github.svg" alt="Measured on agentmods" height="20"></a>Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.
<a href="https://agentmods.dev/skills/sanity-io/agent-toolkit/sanity-form-component-suspense"><img src="https://agentmods.dev/badge/skills/sanity-io/agent-toolkit/sanity-form-component-suspense.svg" alt="Reviewed on agentmods" width="80" height="20"></a>What it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5.1 | $0.00149 | $0.01276 |
| Opus 5.5 | $0.00060 | $0.00510 |
| Sonnet 5.5 | $0.00030 | $0.00255 |
| Haiku 4.5 | $0.00015 | $0.00128 |
Grade A, and why
sanity-form-component-suspense scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured today.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
How it starts
The opening of the file, as written. The whole thing — 88 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Custom form components that suspend
Studio puts a fallback boundary around every field and every array item, so one suspending
component cannot blank the form. That fallback is a generic skeleton: it does not know the size of
what it replaces, so the form shifts when the component loads, and Studio warns about it in the
console. The fix always lives in your code: declare a Suspense around the part that suspends,
with a fallback that takes the same space as what it stands in for. Any boundary you declare is
closer to what suspends than Studio's, so it always wins and nothing warns. That includes a
boundary in a custom field around props.renderDefault(props), which also covers the field's
input: use it when one fallback should cover both, for example a lazy input that also fetches.
Workflow
- Find the component. The warning names the custom components declared for that field or
item (
components.field,.inputor.item), the path, the document type and the schema type, and logs the components themselves; click one in the browser console to open its source. Production builds minify the names, so use the path and document type there. When the schema type declares none, the warning starts with "A form component" and the culprit is usually a plugin'sform.components. React DevTools' Suspense tab shows which component actually suspended. - Find what suspends. Usually one of:
React.lazy,use(promise), an App SDK hook (useDocumentProjection,useDocument,useQuery,useDocuments), or a library in suspense mode (useSuspenseQuery). - Move it into its own component if it is not already, and wrap only that component in
<Suspense>inside the custom component. Keepprops.renderDefault(props)outside the boundary, so the built-in input, field or item never unmounts (and never loses focus while someone types) when the async part suspends again. - Reserve the space. The fallback must have the size of the loaded UI:
- Additive UI (warnings, previews, counts): render it after
renderDefaultso it never pushes the rest down, and give the fallback the height of one loaded row. - A whole component that loads lazily: a skeleton with the component's height, not
null. - Array items: the list is virtualized and measures each row, so an item fallback that is shorter or taller than the loaded row makes the rows below it jump. Match the row height.
- Additive UI (warnings, previews, counts): render it after
- Do not suspend on a document that does not exist yet. A new document only exists once the
first edit creates it, and SDK hooks that read the current document wait for it forever. Gate
on
useFormValue(['_createdAt']): it is unset until the document exists. - Keep promises stable. A promise created during render, in
useStateor inuseMemois thrown away while the component suspends on mount, souse()gets a new one every retry and never settles. Use the SDK hooks, a module-scope cache keyed by the inputs, or a data library. - Verify. Open a new document of the type and an existing one: the warning is gone, the rest of the form is editable right away, and nothing moves when the async part lands.
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- today First seen · 88 lines · 149 tokens per session scan A 7e7e3c3b406f
sanity-form-component-suspense is a skill published in the GitHub repository sanity-io/agent-toolkit (192 stars, last pushed today), licensed MIT. It adds 149 tokens to every session and 1,276 once invoked, about $0.0006 per session on Opus 5.5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-10-23.
Other skills, from other repositories
troubleshooting-assistant
Diagnoses and resolves MCP server registration failures, GPU detection, BigQuery authentication, index build failures, import errors, search quality issues, and performance problems.
opik-diagnose
Surface the Opik traces worth a developer's attention, ranked by signal — Diagnostics issues first, then errors, failed tool calls, latency, regressions, and low online-eval scores. With the Opik MCP connected it lists the project's agentinsightsissue entities, offers to turn Diagnostics on when the project has it…
doctor
Health probe across the worker, embedder, summarizer, queue, vector index, schema, cross-AI wiring and plugin install (cache and live versions). Use when the user types /captain-memo:doctor or when something seems wrong.
inventory
Codebase inventory skill — scans the repo for a named pattern (inline SQL, fetch calls, feature flags, debug leftovers, deprecated APIs, etc.) defined by a YAML "recipe pack" and emits a structured Markdown inventory cross-referenced with framework metadata (routes, handlers). The engine is language-agnostic and…
debug
Read-only bug investigation skill. Takes a bug description + optional stack trace + optional repro steps; orients via code-explorer (Haiku grep/excerpt), then dispatches the debugger persona to emit N ranked hypotheses (1/3/5 on --quick/--normal/--deep) with file:line evidence, verification steps, and confidence…
react-performance
React and Next.js performance optimization patterns adapted from Vercel Engineering's React Best Practices (https://github.com/vercel-labs/agent-skills). Organizes 70+ rules across 8 priority categories — waterfalls, bundle size, server-side, client fetching, re-render, rendering, JS micro-perf, advanced. Use when…