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 kapa-ai/kapa-skills --skill agent-sdk-integrationgit clone --depth 1 https://github.com/kapa-ai/kapa-skillsWrote 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/kapa-ai/kapa-skills/agent-sdk-integration)<a href="https://agentmods.dev/skills/kapa-ai/kapa-skills/agent-sdk-integration"><img src="https://agentmods.dev/badge/skills/kapa-ai/kapa-skills/agent-sdk-integration/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/kapa-ai/kapa-skills/agent-sdk-integration"><img src="https://agentmods.dev/badge/skills/kapa-ai/kapa-skills/agent-sdk-integration.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 | $0.00196 | $0.03481 |
| Opus 5 | $0.00098 | $0.01741 |
| Sonnet 5 | $0.00039 | $0.00696 |
| Haiku 4.5 | $0.00020 | $0.00348 |
Grade A, and why
agent-sdk-integration 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 11d ago.
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 — 133 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Kapa Agent SDK integration
Guide a developer through adding the Kapa Agent SDK to their application, improving an existing integration, or migrating from the Kapa Chat SDK. The Agent SDK embeds an AI agent chat that answers from the customer's knowledge base out of the box (built-in server-side knowledge base search) and can call tools that run in the user's browser.
This skill is a collaborative workflow, not a code generator. Propose, explain the trade-offs, and confirm before making changes. The developer knows their product and their users better than you do.
Golden rules
- Work interactively. At every decision point (which SDK, which tools, approval, history, UI, how the panel opens), present options with a recommendation and a one-line rationale, then let the developer choose. Grouping a few tightly-related quick decisions into one prompt is fine; do not overwhelm with a long chain, and always give concrete defaults.
- Ground every API detail in the source of truth. The SDK evolves. Do not rely on memory for prop names, tool shapes, or endpoints. Before writing integration code:
- Ground against the version you will actually ship, and prefer the latest. For a new setup or a migration, install the current release explicitly (
npm install @kapaai/agent-react@latest, or the repo's package manager). Do this even when a copy already exists innode_modules: a pre-existing install can be stale and older than the current release, and grounding against it produces outdated code. After installing, read that version's types. For an existing-setup review, ground against the version the repo already ships, and see the Improve section for when to suggest an upgrade. - Read the installed package types. Find each package's type entry from its own
package.json(thetypesfield, or thetypesentries underexports). It is commonlydist/index.d.ts,dist/index.d.mts, ordist/index.d.cts, so do not assume a.d.tsextension. Read both@kapaai/agent-reactand@kapaai/agent-core. - Consult the official docs at
https://docs.kapa.ai/dev/agent(fetch them if you have web access). See "Reference links" below. - If the installed types disagree with this skill or the docs, trust the installed types for the version you are shipping. This skill's embedded details track the 1.0.0 API.
- Ground against the version you will actually ship, and prefer the latest. For a new setup or a migration, install the current release explicitly (
- Never invent endpoints, props, or tool APIs. Only wire up endpoints you have actually found in the codebase. Only pass props that exist in the installed types.
- Never expose the API key to the browser. Session tokens are minted server-side. See
references/auth-and-history.md. - Keep the tool set small and high-signal. Fewer reliable tools beat a large grab-bag. Start with the essential ones and expand later.
- Make no destructive changes without buy-in. Show diffs or plans first.
What ships with it
5 files beside SKILL.md in the same directory: the scripts, references and assets a skill reads on demand. Not counted in the per-session cost; read them before you install if any of them is executable.
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.
- 11d ago First seen · 133 lines · 196 tokens per session scan A 105e6375a541
agent-sdk-integration is a skill published in the GitHub repository kapa-ai/kapa-skills (10 stars, last pushed 2mo ago), licensed MIT. It adds 196 tokens to every session and 3,481 once invoked, about $0.0010 per session on Opus 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 2026-08-31.
Other skills, from other repositories
visual-ralph
Visual Ralph orchestration for frontend UI from generated references, static references, or live URL targets, using $ralph with built-in visual verdict and pixel-diff evidence until the implementation matches and leaves a reproducible design system.
debug-optimize-lcp
Guides debugging and optimizing Largest Contentful Paint (LCP) using Chrome DevTools MCP tools. Use this skill whenever the user asks about LCP performance, slow page loads, Core Web Vitals optimization, or wants to understand why their page's main content takes too long to appear. Also use when the user mentions…
repro-admin
Reproduce an EmDash admin UI bug. Attach a container, start the demo dev server, drive the admin with agent-browser using the dev-bypass session, and capture the reproduction as screenshots plus a replayable transcript.
create-site
Creates a new Power Pages code site (SPA) using React, Angular, Vue, or Astro. Guides through the full process from initial concept to deployed site: requirements discovery, scaffolding, component planning, design, implementation, validation, and deployment. Use when the user wants to create, build, or scaffold a new…
prototype-web
A clickable, high-fidelity web product prototype with navigation, a hero section, feature cards, steps, social proof, and optional pricing. It is designed to resemble a finished landing page while remaining a prototype.
menu-transitions-rtl
Animate react-horizontal-scrolling-menu scrolling and build right-to-left menus: noPolyfill defaults to true since v8, so transitionDuration (default 500), a custom-easing-function transitionBehavior, and per-call ScrollOptions { duration, boundary } on scrollToItem/scrollNext/scrollPrev are silently ignored unless…