open-projects
Project management CLI + MCP server for AI agents. Register projects once, open them anywhere, sync to S3, auto-init git, and wire into the hasna ecosystem.
README
open-projects
Project management CLI + MCP server for AI agents. Register projects once, open them anywhere, sync to S3, auto-init git, and wire into the hasna ecosystem (todos, mementos, conversations).
Install
bun install -g @hasna/projects
CLI
# Register a project
projects create --name my-app --path /path/to/my-app
projects create --name my-app # uses cwd as path
# List / search
projects list
projects list --status archived
projects list --tags web,ts # AND filter
projects list --json | jq '.[].path' # machine-readable
# Get / status
projects get my-app # auto-detects from cwd
projects get my-app --json
projects status # all projects at a glance
projects status my-app # single project detail
projects context my-app # full agent handoff context
projects where my-app # paths/workdirs across machines
projects recent # recently opened (with relative times)
projects doctor # health-check all projects
projects doctor my-app --fix --dry-run # preview safe repairs with issue codes
projects doctor my-app --fix # auto-repair what can be fixed
projects stale # stale paths, workdirs, tmux sessions/windows
projects cleanup --dry-run # preview safe stale-record cleanup
projects setup-machine # check Bun/tmux/git/gh/aws/config readiness
projects stats # disk / sync totals
projects stats my-app
# Modify
projects update my-app --description "My app" --tags "web,ts"
projects rename my-app my-app-v2 # updates slug + .project.json
projects tag my-app infra cloud # add tags
projects untag my-app cloud # remove a tag
projects archive my-app
projects unarchive my-app
# Open (cd into it)
cd $(projects open my-app) # print path; auto-detects cwd
workon my-app # shell function — actually cd's
penv my-app # load project .env into shell
projects env my-app # print export statements (eval)
projects env my-app --list # list keys only
# Working directories (multi-machine / multi-repo)
projects workdir add my-app /path/to/dir --label backend
projects workdir list my-app
projects workdir generate my-app # write CLAUDE.md + AGENTS.md
projects workdir generate my-app --dry-run
projects workdir remove my-app /path/to/dir
# Import existing directories
projects import /path/to/existing-project
projects import-bulk /path/to/workspace # imports all subdirs
projects clone my-app /new/local/path # pull from S3 to new machine path
# Sync to/from S3
projects update my-app --s3-bucket my-bucket
projects sync my-app
projects sync my-app --direction push
projects sync-all # sync all projects with S3 configured
projects watch my-app # push changes live as you edit
projects sync-log my-app # show sync history
# Schedule auto-sync
projects schedule set --interval daily --direction both
projects schedule status
projects schedule remove
# Cloud sync (SQLite ↔ RDS PostgreSQL)
projects cloud status
projects cloud pull
projects cloud push
# tmux
projects tmux open --name my-app
projects tmux window-status my-app my-app --json # alive/dead/missing window health
projects tmux dead-windows my-app # exited panes / dead windows
projects tmux revive-window my-app my-app # recreates only missing/dead windows
projects tmux revive-window my-app my-app --force # explicit live-window restart
projects tmux revive-window my-app my-app --force --cwd "$PWD" --command "claude"
projects tmux revive-window master worker # preserves active windows for every linked session
# Publish to GitHub
projects publish my-app --org hasnaxyz
projects unpublish my-app
# Git passthrough
projects git my-app status
projects git my-app log --oneline -10
# Shell completion (includes workon + penv functions)
eval "$(projects completion)" # bash
eval "$(projects completion --shell zsh)" # zsh
MCP Server
Add to your Claude config:
{
"mcpServers": {
"open-projects": {
"command": "projects-mcp"
}
}
}
HTTP mode
Long-lived Streamable HTTP transport (stateless, 127.0.0.1 only):
projects-mcp --http # default port 8827
MCP_HTTP=1 MCP_HTTP_PORT=8827 projects-mcp
Endpoints: GET /health → {"status":"ok","name":"projects"}, MCP at POST/GET /mcp.
Available tools
| Tool | Description |
|---|---|
projects_create |
Register a new project. Returns workingDirectory + post_create_actions |
projects_list |
List projects, filter by status and/or tags |
projects_get |
Get project by ID or slug. Returns workingDirectory |
projects_update |
Update project metadata |
projects_archive |
Archive a project |
projects_open |
Get workingDirectory for a project (also tracks last_opened_at) |
projects_sync |
Sync to/from S3 (incremental, by file hash) |
projects_sync_all |
Sync all active projects with S3 configured |
projects_sync_log |
List recent sync history |
projects_link |
Store integration IDs (todos, mementos, conversations, files) |
projects_import |
Import an existing directory as a project |
projects_import_bulk |
Import all subdirectories of a path |
projects_publish |
Create GitHub repo, add remote, push |
projects_workdir_add |
Add a working directory; optionally generate CLAUDE.md + AGENTS.md |
projects_workdir_list |
List all working directories for a project |
projects_workdir_generate |
Generate CLAUDE.md + AGENTS.md in working directories |
projects_schedule_set |
Enable cron-based auto-sync |
projects_schedule_status |
Get current schedule config |
projects_cloud_status |
Show RDS connection health |
projects_cloud_pull |
Pull from cloud PostgreSQL to local SQLite |
projects_cloud_push |
Push local SQLite to cloud PostgreSQL |
projects_context |
Full agent handoff context for one project |
projects_where |
Workdir locations across machines with existence flags |
projects_setup_machine |
Machine readiness checks for Bun, tmux, git, gh, AWS, paths |
projects_stale |
Find stale paths, workdirs, orphan tmux sessions, and dead windows |
projects_cleanup |
Preview/apply safe stale workdir cleanup |
projects_tmux_window_status |
Check tmux window alive/dead/missing state |
projects_tmux_dead_windows |
List dead tmux windows |
projects_tmux_revive_window |
Safely recreate missing/dead tmux windows |
projects_create response
{
"id": "prj_abc123",
"name": "my-app",
"path": "/path/to/my-app",
"workingDirectory": "/path/to/my-app",
"instruction": "Project created. workingDirectory: /path/to/my-app",
"post_create_actions": [
{
"description": "Register with open-todos for task tracking",
"tool": "mcp__todos__create_project",
"args": { "name": "my-app", "path": "/path/to/my-app" },
"on_complete": "Call projects_link with todos_project_id=<returned id>"
}
]
}
Architecture
src/
├── cli/
│ ├── index.ts # CLI entry point (commander)
│ └── commands/
│ ├── projects.ts # All project commands
│ └── completion.ts # Shell completion
├── db/
│ ├── database.ts # SQLite init, path resolution
│ ├── schema.ts # Migrations
│ └── projects.ts # CRUD + sync log
├── lib/
│ ├── git.ts # git init, .gitignore, passthrough
│ ├── sync.ts # S3 incremental sync
│ ├── project-context.ts # Agent handoff context
│ ├── setup-machine.ts # Machine readiness checks
│ ├── stale.ts # Stale project/workdir/tmux findings
│ ├── import.ts # Import existing projects
│ ├── github.ts # GitHub publish via gh CLI
│ └── scheduler.ts # Cron auto-sync
├── mcp/
│ └── index.ts # MCP server (all tools)
├── types/
│ └── index.ts # Types + error classes
└── index.ts # SDK exports
Data model
projects — id (prj_ + nanoid12), slug, name, description, status, path, s3_bucket, s3_prefix, git_remote, tags, integrations, created_at, updated_at, synced_at
project_files — project_id, relative_path, size, hash, synced_at
sync_log — project_id, direction, status, files_synced, bytes, error, started_at, completed_at
DB path: ~/.hasna/projects/projects.db (override: HASNA_PROJECTS_DB_PATH)
Ecosystem integration
projects_create returns post_create_actions — an array of MCP calls the agent should make to register the project with:
- open-todos (
mcp__todos__create_project) — task tracking - open-mementos (
mcp__mementos__register_project) — persistent memory - open-conversations (
mcp__conversations__create_space) — team coordination (optional)
After each call, use projects_link to store the returned IDs in the project record and .project.json.
S3 sync
- S3 path:
s3://{bucket}/{prefix}/projects/{project-id}/{relative-path} - Incremental: compares local MD5 vs S3 ETag, skips unchanged files
- Skips files >100MB
- Includes
.git/for full restorability - Writes sync history to
sync_logtable
License
Apache-2.0
Recommended Servers
playwright-mcp
A Model Context Protocol server that enables LLMs to interact with web pages through structured accessibility snapshots without requiring vision models or screenshots.
Magic Component Platform (MCP)
An AI-powered tool that generates modern UI components from natural language descriptions, integrating with popular IDEs to streamline UI development workflow.
Audiense Insights MCP Server
Enables interaction with Audiense Insights accounts via the Model Context Protocol, facilitating the extraction and analysis of marketing insights and audience data including demographics, behavior, and influencer engagement.
VeyraX MCP
Single MCP tool to connect all your favorite tools: Gmail, Calendar and 40 more.
graphlit-mcp-server
The Model Context Protocol (MCP) Server enables integration between MCP clients and the Graphlit service. Ingest anything from Slack to Gmail to podcast feeds, in addition to web crawling, into a Graphlit project - and then retrieve relevant contents from the MCP client.
Kagi MCP Server
An MCP server that integrates Kagi search capabilities with Claude AI, enabling Claude to perform real-time web searches when answering questions that require up-to-date information.
E2B
Using MCP to run code via e2b.
Neon Database
MCP server for interacting with Neon Management API and databases
Qdrant Server
This repository is an example of how to create a MCP server for Qdrant, a vector search engine.
Exa Search
A Model Context Protocol (MCP) server lets AI assistants like Claude use the Exa AI Search API for web searches. This setup allows AI models to get real-time web information in a safe and controlled way.