codecks-cli MCP server
Commit Nov 14, 2026Lets an AI agent manage Codecks.io project boards by creating cards, running standups, and scaffolding game development tasks.
Also known as rangogamedev/codecks-cli
- Stars
- 6
- Tools
- —
- Downloads, 30 days
- —
- Latest version
- —
- Last commit
- Nov 14, 2026
- License
- MIT
What is codecks-cli MCP?
codecks-cli lets an AI agent manage Codecks.io project boards by creating cards, running standups, and scaffolding game development tasks.
Summary generated from the project's README
About codecks-cli MCP server
Managing game development task boards is what codecks-cli enables for AI agents and developers using Codecks.io. This tool acts as a bridge for project management, allowing an agent to handle administrative duties like running standups, performing sprint triage, and scaffolding feature cards. It is designed to help game studios organize work into decks and cards without manual overhead.
Users can set up the server locally using a Python installation and an official Codecks API token with read and write access. The system supports various interfaces including a standalone CLI, a Python API, and an MCP server. While the CLI is optimized for lean context and JSON output, the MCP implementation provides advanced features like snapshot caching, team coordination through claim or delegate actions, and richer editor integrations for tools like Claude Code, Cursor, and Windsurf.
What you can do with it
- 1Automated daily standupsAsk your agent to generate a snapshot of current progress including tasks that are done, in-progress, or blocked.
- 2Scaffold complex game featuresHave the agent create a Hero card that automatically generates related sub-cards for code, design, art, and audio.
- 3Synchronize design documentsInstruct the AI to parse a Game Design Document and create corresponding tasks in your project board.
- 4Coordinate multi-agent workflowsUse the MCP server to allow different agents to claim, release, or delegate tasks to one another.
Written from the project's README. Every use case is checked against a quote from it.
Configuration
CODECKS_ATTACH_ALLOW_DIRS
| Name | Description | |
|---|---|---|
| CODECKS_ATTACH_ALLOW_DIRS | — |
Automated checks
Ran Nov 14, 2026Full report →Not verified yet. A server needs a score of 80, no serious finding, and enough data for us to check. How checks work · What to fix for more points
- Source repository linked and reachablerangogamedev/codecks-cli is public and reachablegithub.com/rangogamedev/codecks-cli ↗Pass
- Package names this repository as its sourceNot checked
- Build provenance (Sigstore attestation)Not checked
- Open source license (OSI)MITspdx.org/licenses/MIT.html ↗Pass
- Commit or release in the last yearLast commit or release 2026-11-11github.com/rangogamedev/codecks-cli ↗Pass
- OpenSSF ScorecardNot checked
- Package name not close to a popular packageNot checked
- No critical advisory in the package or its dependenciesNot checked
- No hidden instructions to the model0 tool descriptions and the README read; no override phrases, credential paths or hidden charactersPass
- Destructive tools are labelledNot checked
- Remote endpoint answers the MCP handshakeNot checked
README
View on GitHubcodecks-cli
Native Journey tools are available experimentally. See the MCP reference for their internal-API limitations and validation status.
Game developers should be making games, not managing task boards.
codecks-cli gives your AI agent project management tools for Codecks.io — standups, feature scaffolding, sprint health, batch operations, and multi-agent coordination. Your agent handles the admin. You focus on the game.
Also works as a standalone CLI and Python API. Zero runtime dependencies.
What is Codecks?
Codecks is a card-based project management tool built for game studios — think Trello meets collectible card games. It organizes work into decks (boards), cards (tasks), and supports effort estimation, milestones, and team workflows. Codecks has an official API (since v2.96) but no CLI — codecks-cli fills that gap.
Agent Quick Start
pip install git+https://github.com/rangogamedev/codecks-cli.git
codecks-cli setup # walks you through tokens — no DevTools needed
codecks-cli agent-init --agent # verify: returns account + project context
Upgrading from 0.5 (Codecks v2.96)
Codecks now has official API tokens, and the old browser-cookie token has stopped working reliably (it is officially retired on 2026-12-31). Upgrading takes two minutes:
- In Codecks, open Your Profile → API Tokens, create a token with read & write access, and copy it (it is shown once; it starts with
cdxut_). - Run
codecks-cli setup, paste the token, and pick the deck new cards should go to. - Delete
CODECKS_REPORT_TOKENandCODECKS_ACCESS_KEYfrom your.env; they are no longer used.
Full migration guide — breaking changes, token best practices, troubleshooting: docs/migration-0.6.md. Summary: 0.6.0 changelog.
Your agent can now use codecks-cli <command> --agent via Bash. No special prompt needed — the CLI outputs stable JSON.
Want a plug-and-play PM skill? Copy examples/skills/pm/SKILL.md to .claude/commands/pm.md and use /pm.
Want MCP tools too? pip install "codecks-cli[mcp] @ git+https://github.com/rangogamedev/codecks-cli.git" — see docs/ai-agent-guide.md.
What Your Agent Can Do
- Run standups and sprint triage
- Create, update, archive, and batch-close cards
- Scaffold Hero cards with Code/Design/Art/Audio sub-cards
- Batch operations via
--ids-only | --stdinpipes (~40 bytes/card) - Sync tasks from a Game Design Document
- Coordinate multiple agents via MCP (claim/release/delegate)
Why CLI-First
The CLI is the recommended default for AI agents:
- 25x leaner context than loading 53 MCP tool schemas
- JSON output with
--agentflag - Pipe-friendly batch workflows (
--ids-only,--stdin,@last) - No MCP dependency needed for routine PM work
MCP adds caching, team coordination, and richer editor integrations when you need them.
Human Quick Start
codecks-cli cards --format table
codecks-cli standup --format table
codecks-cli create "Fix login bug" --deck Backlog
codecks-cli attach <uuid> mockup.png
See docs/cli-reference.md for the full command reference.
Attachments
attach (and create --file, and the attach_files MCP tool) only uploads files that resolve inside an allowed root. Allowed by default: the project root and the current working directory, plus any directory listed in CODECKS_ATTACH_ALLOW_DIRS (os.pathsep-separated — : on Unix, ; on Windows). Entries that are not absolute paths, are not existing directories, or name a whole filesystem root are ignored with a warning.
CODECKS_ATTACH_ALLOW_DIRS=/srv/art:/mnt/renders codecks-cli attach <uuid> /srv/art/hero.png
codecks-cli attach <uuid> hero.png --dry-run # resolved path + size + SHA-256, no upload
Dot-prefixed path components, credential-looking names (*.pem, *.key, id_rsa*, *token*, *secret*), and anything under /etc, /proc, or /sys are always refused. The MCP tool takes the same dry_run flag. Full rules: docs/cli-reference.md.
How It Works
CLI (codecks-cli) ─┐
Python API ─┤── CodecksClient ── Codecks HTTP API
MCP Server ─┘ (40 methods)
All three interfaces wrap the same CodecksClient library. The CLI formats output for terminals, the API returns dicts, and the MCP server adds caching, guardrails, and team coordination.
Features
- Card management — create, update, attach files, archive, delete with filtering by status, priority, deck, owner, milestone, tags, and text search
- Feature scaffolding — Hero cards with linked Code/Design/Art/Audio sub-cards
- Daily standups — done, in-progress, blocked, and in-hand snapshot
- Sprint health — blocked, stale, unassigned, and suggested next cards
- Batch operations — pipe workflows and MCP batch tools (up to 20 cards/call)
- Team coordination — claim/release/delegate across multiple agents (MCP)
- GDD sync — parse a Game Design Document and sync tasks to Codecks
- Snapshot caching — MCP reads in <50ms with selective invalidation
- 1000+ tests — full offline test suite, no live API calls
Setup
Run codecks-cli setup — it walks you through everything interactively. No manual token hunting required.
If you prefer manual setup, see .env.example for the configuration format and docs/ai-agent-guide.md for details on each token.
Documentation
| Document | Contents |
|---|---|
| docs/ai-agent-guide.md | Full AI agent setup, CLI reference, MCP setup, customization |
| docs/migration-0.6.md | Upgrading from 0.5.x: breaking changes, API token best practices, troubleshooting |
| docs/cli-reference.md | CLI command reference and Python API |
| docs/mcp-reference.md | MCP tool inventory, caching, error contract, team coordination |
| examples/ | Setup wizard, PM skill, game-dev agent example |
| AGENTS.md | AI agent instructions (tokens, API pitfalls, known bugs) |
| DEVELOPMENT.md | Architecture, dev setup, testing, release process |
| CONTRIBUTING.md | How to contribute |
| CHANGELOG.md | Version history |
Community
Questions? Ideas? Share your agent setup? Visit Discussions.
Contributing
Contributions welcome! See CONTRIBUTING.md.
License
MIT — see LICENSE.
codecks-cli MCP questions
How do I set up codecks-cli for my agent?
Run the interactive setup command which walks you through the token process. You can then initialize your agent context using the agent-init command.
What can an AI agent do with this tool?
Agents can run standups, perform sprint triage, manage cards (create, update, archive), scaffold feature cards with sub-cards, and coordinate tasks across multiple agents.
How do I use it with MCP?
You can install the MCP version specifically using pip with the [mcp] extra.