Skip to content

Add topic: ai-agent #5115

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Closed
TorstenC wants to merge 1 commit into from
Closed

Add topic: ai-agent #5115

Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
17 changes: 17 additions & 0 deletions topics/ai-agent/index.md
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -0,0 +1,17 @@
---
aliases: agent-harnesses, ai-agent-harness, llm-harness
display_name: Agent Harness
short_description: Runtime infrastructure that connects LLMs to real-world tools, memory, and execution environments.
topic: agent-harness
related:
- ai-agent
- orchestration
- tool-use
- mcp
- llm
---
<!-- markdownlint-disable MD041 -->

An **agent harness** is the runtime layer between a large language model and the outside world. It manages the agentic loop — prompt the model, parse tool calls, execute tools, feed results back — while enforcing permissions, compressing context, and handling error recovery. The term gained traction after Anthropic's Claude Code source was studied, revealing the internal architecture that turns a text-generating LLM into a capable software-engineering agent.

Agent harness projects include full runtime environments (LobeHub, oh-my-agent, HexAgent, Utah), agent harness frameworks for Python and TypeScript (water, Chorus), code-quality agents (desloppify), portable skill systems (cc-harness-skills, oh-my-agent), research surveys (LLM-Agent-Harness-Survey), and autonomous systems with persistent memory and orchestration (Adam, Sibyl). The pattern is characterized by a tool registry, a permission system, context management, and often multi-agent coordination.
Loading