agent-config
A Python CLI tool for deploying and managing AI coding agent configurations — MCP servers, instruction files (AGENTS.md, CLAUDE.md), skills, and runtime memory — across multiple providers and machines.
Keep your configs in version control. Run agent-config deploy to push them to every tool.
Install
pip install agentcfg
Or with uv:
uv tool install agentcfg
Quick Start
# Scaffold a new config directory
agent-config init
# Edit the generated files to match your setup
# (mcp_providers.json, agents.json, instruction_providers.json …)
# Deploy everything
agent-config deploy
Commands
| Command | Description |
|---|---|
agent-config init |
Scaffold a new config repo with starter templates |
agent-config deploy |
Deploy manifest entries (files/dirs) to their destinations |
agent-config pull |
Pull memory files back from deployed locations into the repo |
agent-config sync |
Deploy then run agents sync across all LLM providers |
agent-config mcp deploy |
Write MCP server configs to all registered provider files |
agent-config mcp sync |
Additive MCP sync — add only servers missing from disk |
agent-config mcp show |
Show MCP configuration status per provider |
agent-config instructions deploy |
Copy instruction files (AGENTS.md etc.) to each provider |
agent-config instructions show |
Show instruction file status per provider |
agent-config skills deploy |
Deploy all registered skills to their target agents |
agent-config skills sync |
Deploy only skills not yet installed |
agent-config skills show |
Show skill deployment status |
agent-config providers list |
List registered MCP providers |
agent-config providers add |
Add a new MCP provider to the registry |
agent-config providers remove |
Remove an MCP provider |
agent-config providers set-path |
Update a provider's config file path |
Global options
| Flag | Description |
|---|---|
--config-dir DIR / -C DIR |
Root of your config repo (default: cwd or $AGENT_CONFIG_DIR) |
--version / -V |
Print version and exit |
Filtering and dry-run
Most commands accept --providers p1 p2 … to limit to specific providers,
and --dry-run to preview changes without writing anything.
Config Files
| File | Purpose |
|---|---|
manifest.json |
Maps source files/dirs to their deploy destinations |
agents.json |
MCP server definitions and integration toggles |
mcp_providers.json |
Registry of MCP provider config paths and formats |
instruction_providers.json |
Registry of instruction file destinations per provider |
skills_registry.json |
Registry of skills (local directories or skills-cli packages) |
Prerequisites
| Tool | Purpose |
|---|---|
| Python 3.8+ | Runtime for agent-config |
| uv | Package manager |
| just | Task runner (optional, for justfile recipes) |
| Node.js + npx | Required only for installer: skills-cli manifest entries |
| Git | Required for pull command (git add) and submodules |
Environment Variables
| Variable | Description |
|---|---|
AGENT_CONFIG_DIR |
Override the default config directory (equivalent to --config-dir) |
Supported MCP Providers
Out of the box: VS Code, GitHub Copilot CLI, Claude Desktop, OpenCode, Cursor, Windsurf, JetBrains Junie.
Add custom providers with agent-config providers add.
Cross-Platform
Works on Windows (PowerShell), macOS, and Linux. Path expansion supports both
%USERPROFILE% (Windows) and ~ / $HOME (POSIX).
Release files for agentcfg 0.2.1
For a detailed explanation of source distributions (sdists) and built distributions (wheels), please see the package formats documentation.
Source distribution (sdist)
| File | Size | Uploaded | |
|---|---|---|---|
| agentcfg-0.2.1.tar.gz | 19.1 kB | Details |
Built distribution (wheel)
| File | Interpreter | ABI | Platform | Reset |
|---|---|---|---|---|
| agentcfg-0.2.1-py3-none-any.whl | Python 3 | none | any | Details |
Total release size: 37.2 kB
Release files / agentcfg-0.2.1.tar.gz
| Download URL | agentcfg-0.2.1.tar.gz |
|---|---|
| Size | 19.1 kB |
| Tags | Source |
|
SHA-256 checksum How to use checksums |
ce516f57386c68b9b2abb788e4c585f297a5b099c44020716c98ed7712fdd6f3
|
|
BLAKE2b-256 checksum How to use checksums |
0990da77d24d322612c71e67e83fcc78ef39daa7f6fabd3f9062a49001a02986
|
| Upload date | |
|
Uploaded using Trusted Publishing? What is trusted publishing? |
Yes |
| Uploaded via |
twine/6.1.0 CPython/3.13.12
|
Provenance
Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.
PyPI Publish Attestation
PyPI verified that this artifact, at this checksum, originated from the publisher listed below.
Signed by GitHub Actions, verified by PyPI on May 8, 2026.
Transparency logRelease files / agentcfg-0.2.1-py3-none-any.whl
| Download URL | agentcfg-0.2.1-py3-none-any.whl |
|---|---|
| Size | 18.1 kB |
| Tags | Python 3 |
|
SHA-256 checksum How to use checksums |
418cbefac27bf94695f3f89ceebd6ca64b5509d36d4982c6ce4b975355bebb41
|
|
BLAKE2b-256 checksum How to use checksums |
e765ab9157a8f31cd6f2be2cd5a300b21ee9fe9a78bc5706dde702e0349823e0
|
| Upload date | |
|
Uploaded using Trusted Publishing? What is trusted publishing? |
Yes |
| Uploaded via |
twine/6.1.0 CPython/3.13.12
|
Provenance
Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.
PyPI Publish Attestation
PyPI verified that this artifact, at this checksum, originated from the publisher listed below.
Signed by GitHub Actions, verified by PyPI on May 8, 2026.
Transparency log