Open Source Radar — September 19: agents get a spec layer

Share
Open Source Radar — September 19: agents get a spec layer

Today's daily trending board is three-quarters skills and plugins, and the three things that actually broke through are all fixes for the same complaint: the agent does the work, but nothing holds the plan, the context, or the tooling in place. One project makes specs the unit of work, one makes memory a service you can self-host, and one hands an agent your trading charts.


OpenSpec (TypeScript, ~62,000 stars, MIT) — Spec-driven development, packaged as a framework rather than a manifesto. Instead of asking a coding assistant to write a feature straight from a prompt, OpenSpec has the agent draft a proposal, requirements with concrete scenarios, a technical design, and an implementation checklist as plain Markdown in the repo — then you approve the plan before any code lands. It is built for brownfield work, which is the part most spec frameworks skip: the change lives in a folder, gets applied task by task, and is archived with the specs updated, so the repo accumulates a record of what was decided and why. Worth trying on a codebase where the assistant keeps "helpfully" rewriting things you didn't ask it to touch.


Supermemory (TypeScript, ~29,000 stars, MIT) — A memory and context engine that claims the top spot on LongMemEval, LoCoMo, and ConvoMem, the three benchmarks that matter for agent memory, with 95% recall at 15 retrieved items and a reported 99.4% context reduction. The pitch is the whole stack in one system: it extracts facts from conversations, maintains a user profile, resolves contradictions and expired facts, and serves hybrid search that mixes knowledge-base retrieval with personalized context. It runs fully local, ships connectors for Drive, Gmail, Notion, OneDrive, and GitHub, and handles PDFs, images, and code. If you have been gluing a vector store to a summary table and calling it memory, this is the category doing it properly.


TradingView MCP (JavaScript, ~5,900 stars, license: custom) — An MCP bridge that connects a coding agent to the TradingView Desktop app already running on your machine, over Chrome DevTools Protocol. It reads and drives your own charts — analysis, Pine Script development, repetitive chart work — and the README is unusually blunt about what it is not: not affiliated with TradingView, not a paywall bypass, requires a valid subscription and the desktop app, and it does not place trades. Because it drives undocumented internal interfaces, a TradingView update can break it without warning. That caveat is the honest part, and it is also why the fork count is nearly half the star count — people are already building on it.


Worth watching this week: whether OpenSpec's artifact-first flow survives contact with real deadlines, and whether Supermemory's benchmark numbers hold up outside its own harness.

Do you want your agent to plan before it codes, or do you find the spec layer just slows it down? Tell us in the comments.

Sources: OpenSpec (GitHub) · Supermemory (GitHub) · TradingView MCP (GitHub)

Read more

Akhetonics says its all-optical CPU reaches a customer in 2026

Akhetonics says its all-optical CPU reaches a customer in 2026

Light-based computing keeps promising more than it delivers — but one Munich startup has just put a date on its bet, and the interview laying it out is doing the rounds on Hacker News this week. Akhetonics says it will deploy its first commercial machine with a major customer by the end of 2026, with several more planned for 2027. The company, founded by Michael Kissner and Leonardo Del Bino, is building a computer where data enters as light, is switched as light, and circulates through memory

The Week in AI — October 5–11, 2026

The Week in AI — October 5–11, 2026

Every big claim this week turned out to rest on fine print more interesting than the headline: revenue only the company reporting it can define, safety tests sandboxed while the product keeps the web, and a Pentagon phase-out nobody would confirm until reporters kept asking. The week's top 5 1. OpenAI's revenue was $20 billion below the numbers everyone quoted — and the gap was definitional. The Financial Times reported Thursday that OpenAI's annualized revenue runs roughly $20 billion unde

Drone strike shuts a third Yandex data center, taking YandexGPT offline

Drone strike shuts a third Yandex data center, taking YandexGPT offline

Russia's largest tech company is learning what the AI era's infrastructure war looks like from the receiving end — three data centers in four days, and with them much of the cloud layer Russian businesses run on. A Ukrainian drone strike knocked out Yandex's data center in Vladimir early Sunday morning, the third of the company's facilities hit since October 8. The site — reported at roughly 50 MW and designed for about 2,880 server racks — stopped operating completely after the attack, Yandex

Agent teams cost up to 5x more, barely score higher

Agent teams cost up to 5x more, barely score higher

The multi-agent hype train hit a benchmark this weekend — and the grid and the trucking regulators had quiet weeks of their own. Vals AI put agent teams head-to-head with single agents on its Vibe Code Bench, and the teams cost between 1.8 and 5.1 times more for almost no extra quality. The evals company ran GPT-6 Sol and Claude Opus 5.5 solo and in teams across 50 apps at two reasoning efforts; out of four comparisons, only one was statistically significant — Sol at medium effort, where the t