Skip to content

Agent Integration

chant integrates with AI coding agents through three surfaces: skills (operational knowledge), MCP tools (programmatic actions), and Ops (durable deployment workflows). You write TypeScript; the agent uses these surfaces to understand and execute the build → validate → deploy → troubleshoot workflow.

Skills are operational playbooks — structured markdown documents contributed by lexicons. They teach agents how to work with a specific infrastructure platform: which commands to run, what to validate, how to deploy, and how to troubleshoot failures.

Skills are installed by chant init and refreshed by chant update. They live in the project root at skills/:

my-project/
skills/
chant-aws/
SKILL.md
chant-aws-eks/
SKILL.md
src/
index.ts

Agents that support file-based skill discovery load these automatically.

You don’t need to read or manage skill files yourself — they’re installed once and the agent reads them as needed.

The MCP (Model Context Protocol) server gives agents programmatic access to chant operations. When an agent connects via MCP, it can build projects, run lint checks, scaffold files, and more — without the user typing commands.

chant init configures MCP automatically by writing to your IDE’s global MCP config file.

To configure manually, add the chant MCP server to your IDE’s MCP config:

{
"mcpServers": {
"chant": {
"command": "npx",
"args": ["chant", "serve", "mcp"]
}
}
}
MCP toolDescription
buildSynthesize and write infrastructure output
lintRun lint checks on TypeScript intent definitions
importImport infrastructure templates and generate TypeScript files
explainSummarize project resources by lexicon and kind
scaffoldGenerate starter files from lexicon templates
searchSearch available resource types across lexicons
MCP resourceDescription
chant://contextLexicon-specific instructions and patterns for chant development
chant://examples/listList of available chant examples

Lexicon plugins can contribute additional tools and resources.

Ops are named, phased Temporal workflows defined in *.op.ts files. They run durable deployments — phases, gate steps that pause for human signals, compensation on failure — and surface progress in the Temporal UI.

Terminal window
# Start an Op workflow
chant run alb-deploy
# Send a signal to unblock a gate step
chant run signal alb-deploy gate-dns-delegation

The MCP server exposes Op tools (op-list, op-run, op-status, op-signal) so agents can start and monitor deployments programmatically.

See the Ops guide for details.

SurfaceWhenExample
SkillsAlways-on guidance — installed once, agent reads as neededAgent checks the AWS skill before deploying a CloudFormation stack
MCP toolsProgrammatic actions during executionAgent calls build to synthesize templates, lint to validate, op-run to start a deployment
OpsDurable multi-step workflows with observability and gate stepsDeploying infra + app layers, waiting for DNS delegation, rollback on failure

Skills and MCP tools are complementary: skills tell the agent what to do, MCP tools let it do it. Ops tie both together into repeatable, observable workflows.