supercli Documentation

Start here

Getting Started

Run `npx supercli help` for instant usage, install dependencies with `npm install`, copy `.env.example`, and start the local server via `npm start` or `supercli --server`. The web UI (EJS + Vue + Tailwind) is reachable at `http://localhost:3000`. Refer to the repository README for additional context.

Concept

Skill Layer Basics

Every capability becomes a skill, whether it originates from CLI commands, OpenAPI endpoints, MCP tools, workflows, or HTTP requests. supercli indexes names like `namespace.resource.action`, exposes schemas, and lets agents discover, inspect, and execute consistently.

supercli skills list
supercli skills search "database"
supercli skills get beads.issue.create

Sources

Skill Sources

Bundled Harnesses – beads, gwc, commiat ship with the CLI.
Built-in Adapters – OpenAPI specs, raw HTTP, and MCP servers become skills automatically.
Plugins – Install docker, gh, aws, kubectl, terraform harnesses via `supercli plugins install`.
AI Plans / Workflows – Natural language `ask` and `plan` commands create DAGs of skills.

Runtime

Architecture

Agents route requests through the skill discovery layer, which maps namespaces to harnesses/adapters. Execution envelopes guarantee deterministic JSON responses and exit codes.

Agents/Humans → supercli runtime → Skill Router → Harness / Adapter → Tool

Interface

CLI Usage

supercli <harness>                # list skills
supercli beads issue list --json
supercli skills search --harness beads "task"
supercli ask "deploy staging"
supercli plugins explore --tags git,ai

`--json`, `--human`, and `--compact` control output, while exit codes communicate validation, resource, or integration errors.

Extension

Plugins & Harnesses

Wrap any CLI with a `plugin.json` manifest that defines commands, schemas, dependency checks, and passthrough rules. Install from the registry, Git repos, or local paths.

Need more? Check `docs/plugin-harness-guide.md` in the repository for step-by-step authoring details.

Vision

Skill Mesh Direction

The platform is evolving toward a mesh that handles discovery, routing, execution, composition, and governance. Key areas under exploration include a unified skill registry, graph-native discovery, mesh-wide policy, DAG orchestration, agent endpoints, composition tooling, telemetry, and richer plugin UX.

Resources

supercli Knowledge Base