Use Cases
Code agent
Build a coding or code-review agent with filesystem tools, repository context, multi-step runtime behavior, and human oversight.
Code agents need more than completions. They usually need repository access, iteration, memory, and some form of review or approval loop.
Typical stack
npm install @agentskit/adapters @agentskit/runtime @agentskit/tools @agentskit/skills @agentskit/observability @agentskit/evalRecommended package mix
| Layer | Package | Why it matters |
|---|---|---|
| Provider | @agentskit/adapters | Choose the best coding model without rewriting the app |
| Runtime | @agentskit/runtime | Lets the agent inspect, modify, and reflect across multiple steps |
| Tools | @agentskit/tools | Filesystem, shell, browser, MCP, integrations |
| Skills | @agentskit/skills | Useful starting personas like reviewer, planner, researcher |
| Ops | @agentskit/observability | Trace what happened before trusting automation |
| Quality | @agentskit/eval | Replay and benchmark prompts or review behavior over time |
What the architecture usually looks like
- The agent receives a coding or review task.
- Filesystem and shell tools inspect the repo state.
- The runtime loops through investigation, edits, and validation.
- Skills shape behavior for review, planning, or implementation tone.
- Observability and evals help you keep the system trustworthy.
Good defaults
- Separate read tools from write tools when you first ship.
- Add confirmation gates before destructive actions.
- Store structured traces so you can replay bad runs.
- Treat evals as part of the product, not an afterthought.
Best follow-up guides
When AgentsKit is especially strong here
AgentsKit is a good fit when you want to combine tool use, structured runtime control, and post-run inspection without taking on a huge framework. The contracts also make it easier to split the system into safe layers as the agent becomes more capable.