skills.rest
LinXueyuanStdio
LinXueyuanStdioCOMMUNITY
Explore all skills by this author →
WebsiteGitHub

agent-chat

Run non-interactive agent conversations with memory, MCP, and skills loaded.

2|Updated Jan 30, 2026
Search Tags:#agent chat#session resume#mcp integration#non-interactive cli#memory injection#skills orchestration#json streaming
One-click install
npx skills add https://github.com/LinXueyuanStdio/viben --skill agent-chat-linxueyuanstdio
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: agent-chat
Source: https://github.com/LinXueyuanStdio/viben/tree/main/docs/specs/modules/cli
Command: npx skills add https://github.com/LinXueyuanStdio/viben --skill agent-chat-linxueyuanstdio

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

This Skill enables running a configured Viben Agent in a non-interactive way, so you can programmatically ask it questions or trigger coding workflows while it automatically loads memory, MCP settings, and enabled Skills.

Core Features & Use Cases

  • Non-interactive agent dialogue: Send a prompt to a specific Agent ID without opening an interactive chat UI.
  • Automatic memory injection: Loads MEMORY.md plus recent daily logs (today/yesterday) into the system prompt, unless disabled.
  • Session management for continuity: Supports specifying a session ID, resuming an existing session, or starting a new one.
  • Executor selection by agent type: Uses the Agent’s configured type (e.g., claude-code, gemini, codex) to select the correct underlying executor.
  • MCP integration and skills loading: Automatically applies the Agent’s MCP server configuration and enabled Skills.
  • Flexible IO formats: Output and input can be either plain text or stream-json, plus optional JSON result mode.

Quick Start

Ask the specified agent to analyze your code by running: viben agent chat -n my-agent -p "Analyze this repository and propose concrete improvements."

Frequently Asked Questions about agent-chat

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

FAQPage Schema
How do I run an agent non-interactively for automated coding workflows?▼

To run an agent non-interactively, you can pass a prompt via command line or stdin. This automatically loads the agent's memory, MCP configuration, and enabled skills to execute scripted coding assistance without opening an interactive UI.

Can I resume a previous agent session when sending batch prompts?▼

Yes, you can resume a previous agent session when sending batch prompts. The skill supports session management by allowing you to specify a session ID, resume an existing session, or start a new one for continuity.

Does non-interactive agent chat support memory injection from logs?▼

Non-interactive agent chat supports memory injection by automatically loading MEMORY.md and recent daily logs from today and yesterday into the system prompt, unless explicitly disabled during execution.

How does executor selection work for different agent types in CI-like workflows?▼

Executor selection for CI-like workflows works by using the agent's configured type, such as claude-code, gemini, or codex. The skill automatically selects the correct underlying executor based on this configuration.

What IO formats are supported for streaming agent responses in scripts?▼

For streaming agent responses in scripts, both plain text and stream-json IO formats are supported. You can also enable an optional JSON result mode to structure the output programmatically.

Do I need to manually configure MCP servers for non-interactive agent dialogue?▼

You do not need to manually configure MCP servers for non-interactive agent dialogue. The skill automatically applies the agent's existing MCP server configuration and loads enabled skills during execution.

Related Skills

Explore More Software Engineering Skills→
LinXueyuanStdioLinXueyuanStdio
2

viben agent chat

Run non-interactive agent conversations with configured memory and session persistence.

Community
Intermediate
leaderiopleaderiop
7

Packages Graph

Parse package manifests, construct graphs, and generate compliance reports.

Community
Intermediate
c-dalyc-daly

distill

Distill episodic session learnings from EPISODES.md into persistent memory patterns in MEMORY.md.

Community
Intermediate
reggiechan74reggiechan74
5

headless

Run Claude Code non-interactively via CLI with JSON output.

Community
Intermediate