Erik

The Critical Importance of Kindness

[[email protected] @trydotworks]

recursive-mode is a recursive workflow for agentic engineering. As powerful as Factory's Missions, but pre-dates it and is free and open source.

More:
https://t.co/EltmMHGQUN

[[email protected] @trydotworks]

Long-running agent work has a common failure mode: requirements, decisions, and plans live in the conversation. Once the session ends or the context window overflows, the agent loses track of what was decided, what was implemented, and why. https://t.co/EltmMHGQUN

[[email protected] @trydotworks]

recursive-mode solves context rot by making repository documents the source of truth for every phase. Requirements, plans, evidence, state, and decisions live in files that persist across sessions, contributors, and repositories. Prompts become short commands, not full specs.

[[email protected] @trydotworks]

recursive-mode is built on a few simple rules. Each development phase produces one locked output document. Each phase uses earlier phase outputs as input. Before the agent can exit a phase and move forward, it must satisfy defined criteria based on the workflow and prior docs.

[[email protected] @trydotworks]

recursive-mode is a free open source alternative to Missions. It is for developers and teams who want auditable agent runs, recorded evidence, and long or resumable work that can continue without losing context. https://t.co/EltmMHGQUN

[[email protected] @trydotworks]

Before there was Missions by FactoryAI, there was recursive-mode, and it's still free and open source: https://t.co/XHWM4HaeK7

[[email protected] @trydotworks]

Learn how recursive-mode works with this explainer. Note that the diagram is incorrect; recursive-mode performs a self-audit and validation at every phase, not only at the last phase. https://t.co/ezeDFFFaON

[[email protected] @trydotworks]

recursive-mode is an installable skill package for coding agents. It gives your agent a file-backed workflow for requirements, planning, implementation, testing, review, closeout, and memory, instead of leaving the whole process scattered in context.

[Erik @modesvu]

@drakenCSGO bot

[Erik @modesvu]

@jukan05 Nope. We still like Japanese products in China. Depends on the category.

[Erik @modesvu]

@steipete @kevinrose yeah, about a day.

[Erik @modesvu]

@keachhagey why dont you just directly give us the archive link?

[Erik @modesvu]

@mynamebedan people who know software know that openclaw is a dead end. the future direction is skills execution in various types of sandboxes.

[Erik @modesvu]

@teortaxesTex European in China, watching Europe's party functionaries destroy the region for the next couple of decades.

[Erik @modesvu]

@kvncnls Adding functionality while maintaining coherence in the product is another big one. Then on top of that, UI.

[Erik @modesvu]

@0xBOYD We're going to need specialized providers and models that do 1k-10k tks for this - its going to be an interesting year with increasingly specialized models. Never thought UI would require so much power.

[Erik @modesvu]

@poezhao0605 ByteDance's biggest challenge is that their org cannot build international software products, and they cannot build international organizations. Look at the TikTok Shop dumpster fire. ByteDance is going nowhere outside China - TikTok was an acq.

[Erik @modesvu]

@0xBOYD @jpschroeder btw, if you could map out on your homepage or docs how Arrow fits within a stack within a possible stack with AG-UI, AI SDK, ADK, Workers etc and which it replaces/duplicates, that would be very helpful to understand how/if to build on this

[Erik @modesvu]

@0xBOYD @jpschroeder OK, still impressive. Looking forward to more updates.