Claude resources tagged “claude-agent-sdk”
18 curated Claude Code resources tagged claude-agent-sdk in the HeyClaude directory — mostly guides and agents.
Highlights from this set
Standout entries tagged claude-agent-sdk, picked by their own metadata — trust tier, provenance, documentation, and recency.
All claude-agent-sdk resources
Agent SDK Production Architect Agent
Reviews Agent SDK production design — surface tier, allowedTools/permission modes, subagent isolation, session resume, cost from the result message, and sandboxed hosting.
Agent Skills in Claude Agent SDK Applications
Use Agent Skills in the Claude Agent SDK: filesystem discovery via settingSources, the skills option to enable or filter, and tool access.
Building In-Process MCP Tools with the Claude Agent SDK
Define in-process MCP tools for the Claude Agent SDK with createSdkMcpServer and the tool helper, then wire them into query.
Claude Agent SDK Quickstart for Production Agents
Production quickstart for the Claude Agent SDK with bootstrap and guardrails.
Cost Tracking for Claude Agent SDK Applications
Read total_cost_usd from the result message, dedupe parallel tool calls by id, split spend per model, and track cache token fields.
Custom Tools With the Claude Agent SDK
Choose between in-process SDK tools and external MCP servers, with typed schemas and structured error handling.
External Session Storage for Claude Agent SDK Hosts
Mirror Claude Agent SDK session transcripts to S3, Redis, or a database with a SessionStore so any host can resume them.
File Checkpointing in Claude Agent SDK Workflows
Use file checkpointing in the Claude Agent SDK: enable it, capture checkpoint UUIDs, and rewind files by resuming the session.
Hosting The Claude Agent SDK With Multi-Tenant Isolation
Isolate Agent SDK tenants with cwd, config dir, and settingSources options.
MCP With the Claude Agent SDK
Connect MCP servers to Claude Agent SDK apps with stdio, HTTP, and SSE.
OpenTelemetry Observability for Claude Agent SDK Agents
Export OpenTelemetry traces, metrics, and events from the Claude Agent SDK with OTLP, and control what content is exported.
Permission Design for Claude Agent SDK Agents
Least-privilege Agent SDK permissions: the hooks→deny→mode→allow→canUseTool order, allow/deny rules, modes, and a dontAsk lockdown.
Plugins in Claude Agent SDK Deployments
Load local plugins in the Claude Agent SDK via the plugins option, verify them in the init message, and invoke namespaced plugin skills.
Secure Deployment for Claude Agent SDK Applications
Securely deploy Claude Agent SDK apps: isolation choices, least privilege, the proxy credential pattern, and filesystem controls.
Slash Commands in Claude Agent SDK Sessions
Read slash_commands from the SDK init message, dispatch a command via the prompt, and ship custom SKILL.md commands with allowed-tools.
Streaming Output from Claude Agent SDK Agents
Stream real-time output from Claude Agent SDK agents: enable partial messages and read text and tool-call deltas from StreamEvent.
Signal coverage
How these 18 resources score on the trust and safety signals HeyClaude reviews — counted from this set, not the directory as a whole.
A short, calm digest of reviewed Claude resources. Unsubscribe any time.