Agents

An agent is an LLM-driven system component that can reason about a goal and orchestrate multiple steps to achieve it. Agents use Skills to act on the world.

Basically: Agents define how and when it does it. Workflow: User → Agent → Skills → External Systems

Characteristics of Agents

• Goal-oriented
• Capable of planning and sequencing steps
• Chooses which skills to invoke
• Can react to intermediate results
• Often maintains state or memory
• Operates with some level of autonomy

Docs-as-Code Examples

Agent: “Publish updated API documentation” ├─ Skill: Detect changed files ├─ Skill: Validate Markdown and links ├─ Skill: Generate missing sections ├─ Skill: Build documentation └─ Skill: Report errors or success

Agent: “Review this documentation” ├─ Skill: Lint Markdown ├─ Skill: Validate links ├─ Skill: Check style guide compliance └─ Skill: Generate review report

Resources

Conecept:

• AGENTS.md (a simple, open format for guiding coding agents)

How to write good AGENT.md files:

• https://www.humanlayer.dev/blog/writing-a-good-claude-md
• https://github.blog/ai-and-ml/github-copilot/how-to-write-a-great-agents-md-lessons-from-over-2500-repositories/