skills.rest
obispobruno
obispobrunoCOMMUNITY
Explore all skills by this author →
GitHub

agents-build

Extends existing AgentCore agent projects with memory, VPC, multi-agent, browser, and payments capabilities.

1|Updated Aug 1, 2024
Search Tags:#aws#browser automation#multi-agent#agent memory#vpc#agentcore#code interpreter
One-click install
npx skills add https://github.com/obispobruno/dotfiles --skill agents-build-obispobruno
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: agents-build
Source: https://github.com/obispobruno/dotfiles/tree/main/dot_agents/skills/agents-build
Command: npx skills add https://github.com/obispobruno/dotfiles --skill agents-build-obispobruno

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes references (resource) components.

What problem does it solve? Adding capabilities like cross-session memory, VPC networking, browser automation, code execution, or payments to an existing AWS AgentCore agent project requires knowing the right CLI commands, IAM policies, framework-specific wiring code, and deployment caveats. This Skill routes the developer's intent to the correct workflow and generates validated, production-aware code. ## Core Features & Use Cases - Capability Routing: Maps natural-language requests ("remember user preferences", "call from React app", "run pandas in the agent") to one of ten workflow references covering memory, app integration, VPC, multi-agent/A2A, migration, browser, code interpreter, payments, and teardown. - Framework-Specific Code Generation: Produces wiring code tailored to the detected framework (Strands, LangGraph, OpenAI Agents SDK, Google ADK, Claude Agent SDK), including IAM policies and session lifecycle handling. - Pre-flight Validation: Verifies CLI version (>=0.9.0), locates agentcore.json, and validates resource names against CLI limits before generating commands. - Use Case: A developer says "add memory so my agent remembers user preferences" — the Skill reads the project config, helps choose a memory strategy (SEMANTIC, USER_PREFERENCE, etc.), runs agentcore add memory, and generates Strands or LangGraph wiring code that degrades gracefully in local dev. ## Quick Start Ask the agent to add a capability to your existing AgentCore project, for example: "add cross-session memory to my agent" or "add the browser tool so my agent can scrape websites".

Frequently Asked Questions about agents-build

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I add memory to an AgentCore agent?▼

Run `agentcore add memory --name <Name> --strategies SEMANTIC,USER_PREFERENCE`, then wire it into your agent code using the framework-specific pattern (AgentCoreMemorySessionManager for Strands, AgentCoreMemorySaver for LangGraph). Memory is only available after deploy, not during `agentcore dev`.

How do I call a deployed AgentCore agent from my app?▼

Use the bedrock-agentcore SDK's invoke_agent_runtime with the agent's runtime ARN for IAM-authenticated backends, or send HTTPS requests with a JWT bearer token for CUSTOM_JWT auth. Pass a consistent runtimeSessionId to maintain conversation context across calls.

Strands vs LangGraph for AgentCore memory integration?▼

Strands uses AgentCoreMemorySessionManager, which handles retrieval and injection automatically. LangGraph uses the official langgraph-checkpoint-aws package with AgentCoreMemorySaver for checkpointing and AgentCoreMemoryStore for long-term retrieval. Both require the MEMORY_<NAME>_ID env var set after deploy.

Why does memory work after deploy but not locally?▼

The MEMORY_<NAME>_ID environment variable is only injected into the deployed runtime; `agentcore dev` never sets it. Guard your code with an env var check so the agent runs without memory locally, and test memory behavior against a deployed staging target.

Can I migrate an existing Bedrock Agent to AgentCore?▼

Yes, run `agentcore create --type import --agent-id <ID> --agent-alias-id <ALIAS>` to generate a code-first project. Action groups become Gateway targets, but Knowledge Bases and Guardrails are not auto-converted and require manual wiring.

What are the AgentCore resource naming limits?▼

Agent, memory, gateway, credential, and payment resource names allow up to 48 characters with alphanumerics and underscores, starting with a letter. Project names from `agentcore create` are stricter: 23 characters maximum, alphanumeric only, no underscores.

Related Skills

Explore More Software Engineering Skills→
awsaws
2.3k

agents-build

Add memory, VPC networking, and multi-agent orchestration to AgentCore projects.

Official
Intermediate
juanlamadrid20juanlamadrid20

agent-mlops

Automate production deployment and monitoring of AI agents on Databricks.

Community
Advanced
awsaws
2.3k

agents-connect

Wire AgentCore agents to external APIs via Gateway with Cedar policy protection.

Official
Advanced
awsaws
2.3k

agents-debug

Diagnose broken AgentCore agents by analyzing traces, logs, and CLI prerequisites.

Official
Advanced