agenticflow-plugin
By pixelml
Build, deploy, and manage single AI agents or multi-agent workforces as DAG pipelines on AgenticFlow using the af CLI. Create and iterate agents for tasks, attach MCP clients for tools like Google Docs, Sheets, Slack, Notion, GitHub, run orchestrated workflows for research, development, or marketing automation.
npx claudepluginhub pixelml/agenticflow-skill --plugin agenticflow-pluginComponent Overview
Component Details
Skills (3)
Create, run, and iterate on a single AgenticFlow AI agent — one chat endpoint, one assistant, one persona. Use when the user wants a customer-facing bot, a support assistant, a single task agent, or a prompt experiment. Choose this skill over agenticflow-workforce when there's no orchestration between roles (no handoff, no coordinator → workers). Covers `af agent create/update/run/delete`, the `--patch` partial-update pattern for iteration, `af schema agent --field <name>` for nested payload shapes (including suggested_messages, mcp_clients, response_format), the `model_user_config` / `code_execution_tool_config` settings, and safe iteration loops.
Attach external tool providers (Google Docs, Google Sheets, Slack, Notion, GitHub, Apify, etc.) to an AgenticFlow agent via MCP clients. Use when the user wants their agent to read or write external data, call third-party APIs, save outputs to a doc/sheet, or use any tool beyond the model's built-in knowledge. Covers `af mcp-clients list --name-contains`, `af mcp-clients inspect --id` (classify pattern before attach), and the Pipedream vs Composio write-capability distinction — critical for parametric writes. Route traffic through the `af` CLI; the standalone `agenticflow-mcp` server repo lags the CLI and is not recommended.
Deploy and operate a multi-agent AgenticFlow workforce — a DAG of agents that hand off to each other (trigger → coordinator → worker agents → output). Use when the user asks for a team, pipeline, or multi-agent system: research-then-write, triage-then-specialist, dev shop, marketing agency, sales team, content studio, support center, Amazon seller team. Choose this skill over agenticflow-agent when the ask mentions 'team', 'workforce', 'pipeline', 'multiple agents', 'delegation', 'handoff', or names a built-in blueprint. Provides the `af workforce *` command surface, blueprint decisions, graph wiring, MCP attach recipes, and public URL publishing.
README
AgenticFlow AI Toolkit
Connect your AI tools to the AgenticFlow platform.
The toolkit gives your agent (Claude Code, OpenAI Codex, Cursor, Gemini CLI, …) access to AgenticFlow: build AI agents, deploy multi-agent workforces, inspect and attach MCP tool providers, and automate operations via the af CLI.
Install
-
Claude Code — run these in a chat:
/plugin marketplace add PixelML/agenticflow-skill /plugin install agenticflow-plugin@agenticflow-ai-toolkit -
Cursor — install from the Cursor Marketplace (search "AgenticFlow").
-
Gemini CLI — run:
gemini extensions install https://github.com/PixelML/agenticflow-skill -
OpenAI Codex CLI — in Codex, run
/plugins, search AgenticFlow, and select Add to Codex. -
VS Code / other — open the Command Palette and run Chat: Install Plugin From Source, then paste:
https://github.com/PixelML/agenticflow-skill
Prerequisites
The skills assume the af CLI is installed:
# Quick try (no install)
npx @pixelml/agenticflow-cli bootstrap --json
# Global install (recommended for daily use)
npm install -g @pixelml/agenticflow-cli
Requires Node.js 18+.
Get authenticated
- Get an API key. Sign in at app.agenticflow.ai → Settings → API Keys → Create key. Copy the
a9w_...token. - Link the CLI. Either run
af login(interactive — prompts for API key, workspace ID, and project ID; saves to~/.agenticflow/auth.json) OR export env vars in your shell:
Use env vars for CI/CD;export AGENTICFLOW_API_KEY=a9w_xxxxx export AGENTICFLOW_WORKSPACE_ID=<your-workspace-uuid> export AGENTICFLOW_PROJECT_ID=<your-project-ulid>af loginfor a dev machine. - Verify. Run
af doctor --json --strict— should exit 0 withhealth: true. If it fails, thehintin the error envelope will tell you exactly what to fix. - Orient. Run
af bootstrap --json— this is always the first command of every session. It returns auth, agents, workforces, blueprints, playbooks, and a commands cheat-sheet in one call.
What you get
Three flagship skills auto-route based on your prompt:
| Skill | When it fires | What it covers |
|---|---|---|
| agenticflow-workforce | "team", "multi-agent", "pipeline", blueprint names (dev-shop, marketing-agency, amazon-seller, …) | af workforce *, blueprint decisions, graph wiring, publish + public URL |
| agenticflow-agent | "create/build an agent", "support bot", "customer assistant", "system prompt" | af agent *, --patch iteration pattern, schema drilldown |
| agenticflow-mcp | "mcp", "attach tools", provider names (google sheets/docs, notion, slack, gmail, etc.) | af mcp-clients inspect/list/get, Pipedream vs Composio write-safety, attach shape |
Long-form reference material stays in reference/ — CLI setup, business packs, troubleshooting, glossary.
Auto-updates
Plugins update automatically as new capabilities land in the CLI. Major surface changes (e.g. the workforce deploy in v1.6) propagate to the skills once the plugin is refreshed.
Other install methods
If your platform doesn't support plugins, install the skills directly from this repo or use the af CLI standalone:
- Skills (monolithic legacy copy) — single file, still works
- CLI Reference Library — markdown pages usable by any LLM
The standalone MCP server (agenticflow-mcp repo) significantly lags the CLI and is not recommended. Use the af CLI (via npm install -g @pixelml/agenticflow-cli) instead.
Contributing
We don't accept external pull requests. Any pull requests will be automatically closed.
PixelML employees: open pull requests against the upstream source repository rather than this mirror.
License
MIT
Similar Plugins
ai-sdk-agents
Multi-agent orchestration with AI SDK v5 - handoffs, routing, and coordination for any AI provider (OpenAI, Anthropic, Google)
Help us improve
Share bugs, ideas, or general feedback.