DeadDrop - The Inbox for Your AI
Cloud messaging service that connects local AI tools and autonomous agents with external services like Telegram, Slack, Discord, GitHub, and email. Messages accumulate in topics until your local AI retrieves them, enabling asynchronous communication without webhook servers, ngrok tunnels, or always-on infrastructure.
rill - Embeddable, sandboxed scripting to power AI agents
rill is an embeddable, sandboxed scripting language designed for AI agents. It gives LLMs a language they can generate code for and end users can execute without risk. The host application embeds the rill runtime and controls exactly what capabilities are available — no filesystem access, no network access, no side effects unless explicitly provided. The language eliminates null, exceptions, and implicit type coercion at the design level. Dataflow syntax makes code read like the transformation it describes. rill has zero dependencies and runs in Node, Bun, Deno, or the browser.
Claude Code Runner
Command-line tool for executing Claude Code in automated, unattended environments. Runs single prompts, command files, or script sequences in CI/CD pipelines. Features self-correcting loops via runner signals (REPEAT_STEP, BLOCKED, ERROR), real-time tool call streaming, and timestamped session logs. Fresh context per command enables phase-based implementation.
MCP Policy Server
MCP server that enables Claude Code subagents to efficiently access organizational policies and standards without duplication or bloating prompts. Instead of copying full policy documents into agent prompts, teams can reference specific sections using compact notation like §API.7 or §DEPLOY.3.1-5. The server automatically adds only the requested sections to the context when subagents need them.
Claude Code Plugins
Plugin marketplace with three extensions for Claude Code reliability: checkmate runs linters, formatters, and type checkers on every edit; snoop captures session transcripts as JSONL and identifies anti-patterns like loops and scope creep; policies lets teams codify standards once and inject them into subagents using § notation.