By dirmacs
Development stack for AI-assisted multi-repo work. Persistent memory, quality gates, VPS deployment, cross-repo sync.
Track implementation reality vs plan intent with companion documents
Guide for persisting learnings, decisions, and patterns to dstack memory
Cross-repo dependency awareness, sync protocol, and build order for multi-repo work
Pre-commit quality checklist — 5 questions that prevent shipping incomplete work
Use when you want an agent to work autonomously and indefinitely on a long-running project without stalling out. Sets up multiple parallel cron loops on co-prime cadences (3m/7m/11m) that handle execution, quality, and intel work, with a hard anti-idle rule and fallback hierarchy that regenerates tasks from context when the queue empties.
Matches all tools
Hooks run on every tool call, not just specific ones
Executes bash commands
Hook triggers when Bash tool is used
Own this plugin?
Verify ownership to unlock analytics, metadata editing, and a verified badge. GitHub access is read-only (username + org membership).
Sign in to claimOwn this plugin?
Verify ownership to unlock analytics, metadata editing, and a verified badge. GitHub access is read-only (username + org membership).
Sign in to claimBased on adoption, maintenance, documentation, and repository signals. Not a security audit or endorsement.
Development stack for AI-assisted multi-repo work.
Persistent memory, quality gates, VPS deployment, cross-repo sync.
dstack is a Rust CLI + Claude Code plugin that encodes battle-tested workflows from production work across 15+ repos. It gives your AI coding agent persistent memory, awareness of your whole repo fleet, and quality gates that prevent shipping half-done work.
Built by DIRMACS. Born from real pain — Claude degradation incidents, a team communication breakdown, and a client meeting that almost didn't happen.
cargo install dstack
# Or from source
git clone https://github.com/dirmacs/dstack && cd dstack
cargo install --path crates/dstack-cli # or: cargo install dstack
# ~/.config/dstack/config.toml
[memory]
backend = "file" # "file" (default) or "eruka"
[repos]
tracked = ["ares", "pawan", "my-frontend"]
[deploy.ares]
build = "cd ~/projects/ares && cargo build --release"
service = "ares"
smoke = "curl -sf http://localhost:3000/health"
[git]
author_name = "yourname"
author_email = "[email protected]"
dstack config # show configuration
dstack memory load --project myapp # load project memory
dstack memory save "key" "value" # persist a learning
dstack memory query "pattern" # search memory
dstack memory export # export all memory as JSON
dstack sync --status # cross-repo git status (ahead/behind)
dstack sync # pull + push all clean repos
dstack sync --dry-run # preview without pushing
dstack deploy ares # build + restart + smoke test
dstack deploy --all # deploy all configured services
dstack deploy ares --rollback # rollback to previous binary
dstack audit # workspace summary dashboard
dstack audit --pre-commit # quality gate checklist
dstack audit --stale # find stale companion docs
dstack init my-plugin # scaffold a new dstack plugin (6 platforms)
dstack init --dry-run my-plugin # preview files without writing
dstack skills list # list installed skills
dstack skills sync # sync skills from configured repo
| Crate | What |
|---|---|
| dstack | CLI binary + library: memory, deploy, sync, audit, init, skills |
| dstack-memory | MemoryProvider trait, FileProvider (JSON), ErukaProvider (REST API) |
| dstack-server | HTTP API + MCP server mirroring CLI commands |
7 skills, 4 hooks, 3 commands. Works alongside superpowers and GSD — no conflicts.
# Install as Claude Code plugin
claude plugin install /path/to/dstack/plugin
| Skill | Purpose |
|---|---|
| using-dstack | Session orientation, available commands |
| eruka-memory | When/how to persist learnings |
| multi-repo-ops | Cross-repo dependency awareness, build order |
| vps-deploy | Build + restart + smoke test workflow |
| companion-docs | Track implementation reality vs plan intent |
| quality-gates | Pre-commit 5-question checklist |
| ralph-loop | Autonomous iteration via co-prime parallel cron loops — never-idle agent for multi-day projects |
| Hook | Event | What |
|---|---|---|
| session-start | SessionStart | Loads project memory, shows repo status |
| quality-gate | PreToolUse | Injects quality checklist on git commit |
| proprietary-guard | PreToolUse | Blocks commits containing internal/client names in public repos |
| context-monitor | PostToolUse | Warns at 50/100 tool calls to save findings |
| Command | What |
|---|---|
/deploy [service] | Deploy a service or all services |
/sync [status|dry-run] | Sync tracked repos |
/audit [stale] | Quality audit |
File (default) — JSON files at ~/.local/share/dstack/memory/. Zero dependencies. Works offline.
npx claudepluginhub dirmacs/dstack --plugin dstackBootstrap - Developer onboarding accelerator. Analyzes project requirements, detects missing dependencies, generates Docker/docker-compose configs, creates .env templates, sets up pre-commit hooks, configures IDE settings (VSCode/Cursor), and troubleshoots 'it works on my machine' issues. Gets developers productive in minutes, not hours.
114 battle-tested skills for Claude Code. Session memory, git workflows, database migrations, security scanning, deployments, and more. Auto-trigger commands and project handoffs via markdown-based skills.
(Alpha) Persistent memory, architectural decisions, and safety guardrails for Claude Code. Your agent starts every session with full project context — stack, decisions, patterns, safety rules, and a handoff from the previous session.
Claude Code skill pack for Replit (30 skills)
Complete developer toolkit for Claude Code
Git workflow and version control specialist for git strategies (gitflow, trunk-based), merge conflict resolution, interactive rebase, cherry-pick, bisect, reflog, hooks, and submodules. Use when working with complex git operations, resolving conflicts, or implementing git workflows.