| name | AINL |
| description | Deterministic compiled graphs for OpenClaw (importer, runtime, MCP tools) |
| install | ./install.sh |
| install_alt | pip install 'ainativelang[mcp]' && ainl install-mcp --host openclaw |
| commands | ["ainl import <source> — convert Markdown / ecosystem packs into `.ainl` graphs","ainl compile <file.ainl> — compile and validate deterministic graph IR","ainl run <file.ainl> — run compiled workflow (when your graph uses the runtime path)","~/.openclaw/bin/ainl-run <file.ainl> — compile-then-run wrapper after install-mcp --host openclaw (add ~/.openclaw/bin to PATH)"] |
| tags | ["ainl","graphs","mcp","openclaw"] |
| metadata | {"openclaw":{"kind":"skill","upstream":"https://github.com/sbhooley/ainativelang","package_pypi":"ainativelang","mcp_server_key":"ainl","config_file":"~/.openclaw/openclaw.json","workspace_default":"~/.openclaw/workspace","onboarding_note":"OpenClaw expects `npm install` + `openclaw onboard` for the host CLI; this skill adds AINL (pip) + MCP + ainl-run.","install_channels":["ClawHub (when listed)","Manual copy to ~/.openclaw/skills or <workspace>/skills"]}} |
AINL is an AI-native language that compiles deterministic graphs. In OpenClaw, use it to import workflows, compile them, run them, and expose ainl-mcp tools to the agent via ~/.openclaw/openclaw.json.
Package: ainativelang v1.8.0 (this tree; PyPI after publish — docs/RELEASING.md). Optional extras: [mcp], [solana].
Follow this skill when the user is on OpenClaw (or wants deterministic .ainl graphs with import / compile / run / MCP). Prefer ./install.sh from a checked-out skill folder, or the manual pip install 'ainativelang[mcp]' && ainl install-mcp --host openclaw path (alias ainl install-openclaw). OpenClaw does not use zeroclaw skills install <url> — install via ClawHub (if available) or copy this folder into ~/.openclaw/skills or <workspace>/skills. After bootstrap, suggest the morning briefing prompt example below when it fits.
.ainl) instead of ad-hoc prompts.ainl-mcp).From this skill directory (after ClawHub or manual copy): run ./install.sh
That optionally refreshes the OpenClaw CLI via npm, upgrades ainl[mcp], and runs ainl install-mcp --host openclaw (MCP merge into openclaw.json, ~/.openclaw/bin/ainl-run wrapper).
Manual (no skill checkout):
pip install 'ainativelang[mcp]' && ainl install-mcp --host openclaw
| Command | Purpose |
|---|---|
ainl import … | Import Markdown or ecosystem sources into .ainl |
ainl compile … | Compile / validate graphs |
ainl run … | Execute a graph via the CLI runtime where applicable |
~/.openclaw/bin/ainl-run <file.ainl> | After install-mcp --host openclaw: wrapper that compiles then runs |
Tell the user they can say in OpenClaw:
Import the morning briefing using AINL.
(They should point the importer at their Markdown or pack; ainl import subcommands match their source type.)
Configure the host so mcp.servers.ainl in ~/.openclaw/openclaw.json runs ainl-mcp as a stdio MCP server (see AINL docs: OpenClaw integration / External orchestration). ainl install-mcp --host openclaw merges that entry when missing.
openclaw_ainl_gold_standardopenclaw_host_ainl_1_2_8): docs/operations/OPENCLAW_HOST_AINL_1_2_8.md — openclaw_host_ainl_1_2_8ainativelang v1.8.0 (this tree; PyPI after publish — docs/RELEASING.md)