Luciano Castillo

Introducing CRABWISE 🦀 Build trust with your agents and set rules they can't break. Local-first agent monitoring, audit trail, and policy enforcement with multi-agent, multi-provider support → live agent monitoring → local policy enforcement (commandments) → a hash-chained audit trail → terminal UIs for status, watch, and audit → support for OpenClaw, Claude Code, Codex CLI (more to come) + pairs nicely with Crabwalk🦀 Follow me @luccasveg to stay updated + RT!

@dotta Team voice meeting. I need to have an all-hands with my agent team. ASAP.

I'm challenging myself to make sure I setup at least one Paperclip Routine per day that makes Paperclip better Today I'm having it pull my bookmarks and write a report on what we could do to make Paperclip better

@cjc Thanks for taking the time, looking forward to what you build!

Got it, it sounds like linear gathering context ‘globally’ is the feature, as it should be. My big challenge with Linear and other human-native tools is that I struggle to see the value of human -> agent -> human work, outside of enterprise teams. It seems to me that Linear, Jira, etc were built to help humans solve p2p production challenges in product teams. But now with agentic workflows the p2p production team principle is in flux, so how can we assume that the future of agentic product work will be anything like traditional lists and Kanbans? I’d love to see linear ship 10y forward-thinking agentic work prototypes using the talent you have on your teams. Let loose.

The difference is native context. External agents act on Linear and our native agents understand your product and users. Last year, we launched the Linear agent for Slack, Gong, Intercom and Zendesk to bring user feedback directly into Linear, so you have that context in making decisions about what to build next. The natural next step was bringing the agent into the workspace itself. By grounding it in your feedback, docs, and soon your code, we’re moving from tracking issues to helping execute the work, planning projects and triaging at scale. If you prefer using other tools, our APIs and MCP server are only getting better. But for teams wanting a seamless agentic workflow where they actually build, this is the future.

Issue tracking is dead. 🪦 We’ve been trapped in a handoff model where process became the work and engineering time was wasted on ceremony. That era is over. Today, we’re launching the Linear Agent, Skills and Automations. The future is execution, not overhead: linear.app/next

@abacusai Missed out on naming it Clawbacus

OpenClaw now runs without setup, infra, or technical overhead. OpenClaw just got a lot easier to use.↓🧵

@garrytan I have Claude writing plans and Codex reviewing them. Claude has creativity that’s important in planning, Codex is the realist.

I added a /codex skill that does plan and code review in Claude Code and it has found so many more bugs and helped with far more elegant architectural designs And tomorrow I will launch native Codex support so you can use GStack on Codex

In honour of the closure of the metaverse, I would like to re-up one of the worst videos ever recorded.

@0xMartz I’ll take a GitHub star, thanks!

@luccasveg Fire, are you open to accepting any donations via GitHub btw?

I’m building this! a similar guardrail, but one that is multi-agent, multi-provider, not just OpenClaw / NemoClaw crabwise.ai

@0xMartz github.com/crabwise-ai/cr…

@luccasveg Very cool, do you have a GitHub for it?

my gstack fork: github.com/luccast/gt-pro…

I added a Senior Product Designer to @garrytan 's gstack. Now its a product team.

Claude, cancel all my Sunday plans

when building, i now spend 90% of my time thinking and planning about how to optimally deploy autonomous agents to gather context in a way that will help me understand the problem better, so i can orchestrate them better when it's --dangerously-skip-permissions time. it feels like sci-fi

Thou shalt not execute market orders. 📜 x.com/luccasveg/stat…

@SaidAitmbarek thanks! try it out. sure, every bit helps.

@luccasveg dope idea actually love the rule enforcement approach, what i'd expect from a real harness let's launch you on microlaunch.net if that helps.

Introducing CRABWISE 🦀 Build trust with your agents and set rules they can't break. Local-first agent monitoring, audit trail, and policy enforcement with multi-agent, multi-provider support → live agent monitoring → local policy enforcement (commandments) → a hash-chained audit trail → terminal UIs for status, watch, and audit → support for OpenClaw, Claude Code, Codex CLI (more to come) + pairs nicely with Crabwalk🦀 Follow me @luccasveg to stay updated + RT!

@theflohart thanks!

@luccasveg solid UI, and sth i've been buildin on too

@jake_researcher Thanks. One commandment set is enforced across all supported agents on the machine.

@luccasveg The hash-chained audit trail is a smart choice for compliance. How do you handle policy conflicts when multiple agents have different commandment sets?