bytedance/agentkit-samples is a collection of examples and tutorials for Volcengine AgentKit, an AI-agent development platform for building, deploying, and operating agent applications. Developers use the samples to learn agent creation, multi-agent collaboration, memory, retrieval, MCP integrations, media generation, customer service, and other workflows. The catalogue skills provide agent workflows based on these examples.
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 bytedance/agentkit-samples --skill agentkit-hybrid-cloud-demogit clone --depth 1 https://github.com/bytedance/agentkit-samplesWrote 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/bytedance/agentkit-samples/agentkit-hybrid-cloud-demo)<a href="https://agentmods.dev/skills/bytedance/agentkit-samples/agentkit-hybrid-cloud-demo"><img src="https://agentmods.dev/badge/skills/bytedance/agentkit-samples/agentkit-hybrid-cloud-demo/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/bytedance/agentkit-samples/agentkit-hybrid-cloud-demo"><img src="https://agentmods.dev/badge/skills/bytedance/agentkit-samples/agentkit-hybrid-cloud-demo.svg" alt="Reviewed on agentmods" width="80" height="20"></a>- NVIDIA SkillSpector warn
SkillSpector: 3 findings, up to medium
These are SkillSpector’s own severities. On a checked sample its high-severity flags on skills were ~96% false positives — a documented command, a public API, a “never do X” rule — so we show them as a caution to read, not a verdict. Why →
- medium Excessive Agency · line 32 Skill enables autonomous high-impact decisions without human-in-the-loop verification. Critical operations (destructive commands, financial transactions, data deletion) should require explicit user confirmation.Fix: Add human-in-the-loop confirmation for destructive, irreversible, or high-impact operations. Never auto-execute commands that modify files, send data, or alter system state.
- medium Agent Snooping · line 109 Skill enumerates or reads other installed skills. Access to other skills' SKILL.md files or the skills directory reveals prompt instructions, capabilities, and secrets that should be invisible to peer skills.Fix: Remove all code or instructions that list or read other skills' files or directories. Skills should operate independently; cross-skill access is a privilege escalation.
- low Tool Misuse · line 29 Tool parameters are crafted to achieve unintended or unsafe behavior. Parameter abuse can bypass intended safety checks (e.g. shell=True, --force, dangerous glob patterns).Fix: Validate all tool parameters against an allowlist. Reject dangerous parameter values (shell=True, --force, -rf /) and use safe defaults.
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 | $0.00079 | $0.04035 |
| Opus 5 | $0.00039 | $0.02018 |
| Sonnet 5 | $0.00016 | $0.00807 |
| Haiku 4.5 | $0.00008 | $0.00404 |
Grade A, and why
agentkit-hybrid-cloud-demo scanned grade A with 1 finding against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 8d ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Makes network callslowCapability
Not a fault in itself. Listed so you know the mod talks to something, and to what.
- Verify `${AGENTKIT_OPENAPI_SCHEME:-http}://${AGENTKIT_OPENAPI_HOST}/ping` before deployment. This delivered POC defaults to HTTP; honor an explicit HTTPS setting with normal certificate validation. Never make `curl -k` How it starts
The opening of the file, as written. The whole thing — 204 lines — stays where its author put it; the contents beside it link to each section on GitHub.
AgentKit Hybrid Cloud Demo
Operating contract
- Treat
README.mdas the customer-facing source of truth. Read the requested step and its linked document before acting. - Start from a clean validation directory. Preserve unrelated or pre-existing changes.
- Execute steps in README order unless the user names one specific step. Do not silently skip a failed prerequisite.
- Prefer the repository scripts over reconstructed commands:
- Human-guided Base Runtime:
scripts/deploy_interactive.sh - Non-interactive/CI Base Runtime:
scripts/deploy_hybrid.sh - Resource association and release:
scripts/bootstrap_platform.py - Knowledge/Memory proof:
scripts/verify_knowledge_memory.py - Local guide and Chat UI:
scripts/run_local_ui.sh
- Human-guided Base Runtime:
- Perform safe read-only discovery automatically. Stop for genuine human-only actions such as hosts/DNS changes requiring administrator access, console account switching, model selection, uploading documents, obtaining temporary credentials, or production authorization changes.
- Never write or print AK/SK, passwords, Runtime API Keys, model keys, temporary Registry tokens, Bearer tokens, or injected component credentials. Use placeholders in docs and commands.
- After each step, report:
- action and evidence;
- pass, fail, partial, or blocked;
- required human action;
- README/FAQ defect found and the minimal proposed update.
Mandatory safety and validation
- Verify
${AGENTKIT_OPENAPI_SCHEME:-http}://${AGENTKIT_OPENAPI_HOST}/pingbefore deployment. This delivered POC defaults to HTTP; honor an explicit HTTPS setting with normal certificate validation. Never makecurl -kthe customer solution, and clearly distinguish the POC HTTP default from a production HTTPS requirement. - Default customer deployment to
live. For the sample's default Ark profile, requireMODEL_AGENT_API_KEYand resolve the documented default Model Name/API Base. For a custom or real environment, honor explicit API Base, Model Name, and API Key overrides. Inject all three resolved values during launch, fail clearly when the key is absent, and never silently downgrade to demo. UseAGENTKIT_DEPLOY_MODE=demoonly when the user explicitly authorizes an infrastructure-only fallback, and label that result incomplete. - Keep model values out of the public config, project
.env, Git, Docker build context, logs, and prompts. Use the deployment script's transient0600config or a platform Runtime Secret. - Do not manually
docker pullor proactively requiredocker login. Ifagentkit launchactually returnstoken expired,unauthorized,invalid token claims, orauthentication required, stop and explicitly guide the user to refresh the target Registry's temporarydocker login, confirmLogin Succeeded, and rerun launch. This manual login is the required recovery path after an observed authentication failure, not a prohibited action. Never ask the user to paste the token into chat or save it in the repository. - Never silently reuse a global Region for Runtime creation. Require the user to confirm
VOLCENGINE_REGION, preferably throughscripts/deploy_interactive.sh; show any detected global value only as a hint. - Before launch, detect an exact same-name Runtime in the confirmed Region. In an interactive session, ask whether to update that Runtime or enter a new unique name; check the new name before creation and never silently overwrite. In automation, require an explicit
AGENTKIT_RUNTIME_IDor reuse authorization. Persist only the non-secretlaunch_types.hybrid.runtime_nameandruntime_idafter a successful launch so later runs update rather than hitInvalidParameter.DuplicateName; keep endpoint and API/model keys transient. - Treat
scripts/deploy_interactive.shas the single first-deployment entry point for a human. It may reuse a confirmed existing CLI target or collect OpenAPI scheme/host, hidden AK/SK, Region, model profile, and hidden model key interactively. Do not require users to prepare exports or an Agent Prompt for this first deployment. After it succeeds, route later validation and capability work through the roadmap Prompts, which load this project Skill. - Build and publish
linux/amd64, including from Apple Silicon. - Treat
agentkit launchoutput as insufficient proof. Require RuntimeReady, serviceRUNNING, healthHealthy, and a successful/invoke. - Distinguish connection transport from Agent data mode:
transport=remote, mode=demoproves the remote Runtime was called but still uses the deterministic demo; onlytransport=remote, mode=liveproves the live Agent data plane handled the request. - After
runtime update, runruntime releaseand wait forReady/Healthy. - List platform resources with explicitly limited non-sensitive fields. Never dump full Tool JSON or environment variables.
- Do not delete, destroy, reset, or overwrite resources unless the user explicitly authorizes the exact target.
What ships with it
1 file 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.
- 8d ago First seen · 204 lines · 79 tokens per session scan A ee1e54fe811a
agentkit-hybrid-cloud-demo is a skill published in the GitHub repository bytedance/agentkit-samples (450 stars, last pushed 5d ago), licensed Apache-2.0. It adds 79 tokens to every session and 4,035 once invoked, about $0.0004 per session on Opus 5. A static security scan graded it A with 1 finding (makes network calls). No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.
Other skills, from other repositories
claw-admin
Claw system administration: service management, IM connections, logs, cron, and workspace diagnostics. Use when the user asks to manage claw services, connect/disconnect IM platforms, view logs, or perform system-wide operations.
x-nets
Enhanced netstat module with cached data and structured output. View network connections, routing tables, and interface statistics in interactive or TSV/CSV formats. Dependency: This is an x-cmd module. Install x-cmd first (see x-cmd skill for installation options). see x-cmd skill for installation.
install
Install software any way the system allows. Recommends x install — one front door listing every method and picking the best; plus x eget/x env use/x pixi channels and system package managers with mirror switching. All x-cmd channels install in $HOME — no sudo, no pollution. Use for "install", "apt", "brew", "eget"…
x-theme
Change terminal command line prompt theme. Globally manage the display style of all x-cmd interactive components. Supports multiple theme vendors: theme, starship, ohmyposh. Dependency: This is an x-cmd module. Install x-cmd first (see x-cmd skill for installation options). see x-cmd skill for installation.
x-arp
Display ARP cache table with MAC vendor lookup and suspicious entry detection. Supports CSV, TSV, and TUI output. Dependency: This is an x-cmd module. Install x-cmd first (see x-cmd skill for installation options). see x-cmd skill for installation.
x-lsof
Enhanced lsof (List Open Files) with interactive UI and structured output. View open files, network connections, and processes. Dependency: This is an x-cmd module. Install x-cmd first (see x-cmd skill for installation options). see x-cmd skill for installation.