# Build OpenServ AI Agents

Developers need clear guidance to build OpenServ agents without missing platform requirements. This skill explains SDK patterns, provisioning, deployment, troubleshooting, and agent workflows.

## Install

```bash
npx skillstore add internet-court/openserv-agent-sdk
```

## Metadata

- Status: approved
- Slug: internet-court-openserv-agent-sdk
- Skillstore revision: r1
- Version status: missing
- Tree hash: 71898b48419b93968d81623ce81e91d4e208bb14bb6e32645f8efab5e16615ae
- Author: internet-court
- GitHub username: internet-court
- License: MIT
- Repository: https://github.com/internet-court/internet-court-skill/tree/main/vendored/openserv/openserv-agent-sdk
- Ref: 3f6e026a3363e0954ede7bef0cfe88d4475de137
- Supported tools: Claude, Codex, Claude Code
- Audit status: complete
- Agent install advisory: confirmation\_required
- Manual install advisory: allowed
- Artifact signature: available
- Audit attestation: unavailable
- Human verification: not\_verified
- Risk factors: external\_commands, network, env\_access, scripts
- Quality score: 38
- Quality tier: warning
- Public page: https://skillstore.pages.dev/skills/internet-court-openserv-agent-sdk
- Manifest: https://skillstore.pages.dev/api/skills/internet-court-openserv-agent-sdk/manifest

## Capabilities

- Explains how to define OpenServ agents with system prompts and capabilities.
- Shows runless and runnable capability patterns with Zod schemas.
- Documents platform-delegated generation through the SDK generate method.
- Guides provisioning, triggers, local runtime, tunnel behavior, and deployment.
- Covers task, workspace file, payment, and ERC-8004 identity workflows.
- Provides troubleshooting guidance for credentials, ports, tunnels, and module imports.

## Use Cases

- Create a First Agent: Follow the basic agent pattern to provision and run a TypeScript OpenServ agent.
- Deploy Platform Workflows: Prepare agents for webhook, manual, cron, or x402 triggers with correct runtime settings.
- Add On-Chain Identity: Use the ERC-8004 guidance to register an agent identity after provisioning.

## Prompt Templates

### Plan a Basic Agent

```
Help me design a simple OpenServ agent. Include the agent purpose, one capability, required packages, and provisioning steps.
```

### Convert a Tool to a Capability

```
Turn this tool idea into an OpenServ runnable capability. Include input fields, validation needs, task logging, and expected output.
```

### Debug Provisioning Errors

```
Diagnose this OpenServ agent startup error. Explain likely causes, credential checks, and the safest next steps.
```

### Design a Production Workflow

```
Design a production OpenServ workflow with deployment, trigger choice, credential handling, timeout settings, and ERC-8004 considerations.
```

## Limitations

- It is documentation, not a live OpenServ account or hosted runtime.
- It may lag behind current OpenServ SDK releases unless the skill is updated.
- Some workflows require OpenServ credentials, wallet funding, or platform access.
- Users remain responsible for securing API keys, auth tokens, and wallet private keys.

## Best Practices

- Use runless capabilities or generate when platform-delegated model calls are enough.
- Keep user API keys, auth tokens, and wallet private keys out of logs and commits.
- Call provision before deployment and confirm trigger timeouts match expected workloads.

## Anti Patterns

- Do not use the legacy process method unless direct OpenAI calls are required.
- Do not reuse agent API keys for PlatformClient management calls.
- Do not clear wallet state unless you intentionally want a new on-chain identity.

## Security Audit

- Audited at: 2026-07-21T04:24:39.859\+00:00
- Summary: This is a documentation-only skill teaching how to build agents with the OpenServ TypeScript SDK \(@openserv-labs/sdk\); it ships Markdown reference docs and .ts example files, not an executable installer. Of 248 static hits, 237 are false positives: the "shell backtick execution" flags are JavaScript template literals and Markdown code fences, and the "API/secret key" and ".env" flags are placeholder names in an env.example and configuration prose with no real secrets embedded. 11 findings are kept at low severity for visibility because the documented provision\(\) flow persists a crypto WALLET\_PRIVATE\_KEY to .env for on-chain \(ERC-8004 / x402 / USDC\) payments — a transparent, documented behavior a reviewer should be aware of, not exfiltration.

## Stats

- Views: 0
- Downloads: 4
- Favorites: 1
- Popularity score: 0